The Australian Signals Directorate (ASD) has released new guidance on Agentic AI Harnesses, highlighting the security, governance and operational risks organisations need to consider as they adopt agentic AI systems. The publication shifts attention beyond large language models (LLMs) to the software layer that connects models with organisational data, tools and systems. According to ASD, the AI harness is a critical component of an agentic AI system because it determines what information an agent receives, which tools it can access, what actions it can perform and what controls are applied.

Why Agentic AI Harnesses Matter

An agentic AI system combines an LLM with external tools, data sources, memory and planning workflows. While the LLM processes information and proposes actions, the harness provides the surrounding infrastructure needed to execute those actions and enforce permissions. ASD said many significant organisational risks emerge when an LLM is connected to enterprise systems through a harness. Security, privacy, governance and operational risks can depend on what an agent can access, the actions it can perform and how it interacts with connected systems. The publication also notes that some risks, including prompt injection, cannot be reliably addressed within the model alone. Additional controls are required across the harness, connected systems and organisational governance processes.

Harness Becomes a Long-Term Security Focus

ASD describes the harness as the component organisations can most directly govern, secure and control. While LLMs can be replaced as models evolve, the harness and its surrounding integrations are expected to represent a longer-term organisational investment. The harness can manage context and memory, provide access to tools, enforce execution privileges and record activity for monitoring and evaluation. Its components can include a user interface, policy layer, context manager, model interface, tool registry, permission system, execution environment, connector layer, memory and session store, and audit and observability capabilities. [caption id="attachment_114032" align="aligncenter" width="600"] Image Source: ASD- https://www.cyber.gov.au/[/caption] This makes the harness a key configuration surface for agentic AI security, particularly where organisations need to control tools, permissions, approvals and system integrations.

ASD Highlights Key Agentic AI Risks

The guidance identifies five broad risk categories associated with agentic AI systems: privilege, design and configuration, behavioural, structural and accountability risks. Excessive privileges can allow a compromised agent to have far-reaching access, while insecure architecture or configuration can introduce weaknesses before deployment. Behavioural risks include unintended actions and manipulation through techniques such as prompt injection and data poisoning. Structural risks can emerge when failures cascade across interconnected components and workflows. Accountability risks may arise when complex agentic systems make it difficult to trace decisions, audit actions or assign responsibility. ASD also recommends treating multi-agent systems as a single agent for security purposes because a compromise in one component may propagate through shared context and trust relationships.

Security Controls for Agentic AI

The publication recommends established cybersecurity practices, including least-privilege access, identity and access management, secure design, monitoring and incident response. Organisations are also advised to require human oversight for high-impact actions, control access to external data and tools, validate agent outputs, maintain audit logs and apply supply-chain assurance. ASD recommends a phased approach to deployment, beginning with approved use cases, appropriate data classification and security validation before broader adoption. The guidance also stresses that no harness is inherently secure. Organisations should select harnesses appropriate to their tasks, understand their capabilities and permissions, and apply controls across multiple layers rather than relying solely on model behaviour or prompts. For executives and CISOs, ASD recommends asking what data and systems an agent can access, what actions require human approval, how AI-specific attacks are mitigated, whether significant actions can be monitored and audited, and what could happen if the harness were compromised or misconfigured. As agentic AI adoption expands, ASD's guidance positions the harness as a central part of managing the technology's security, governance and operational risks.

Some of the world's leading democracies are pushing governments and companies to start preparing for post-quantum cryptography before quantum computers become powerful enough to break the encryption systems that protect global digital infrastructure today.

In a joint advisory released Thursday, the G7 Cybersecurity Working Group and the U.S. Cybersecurity and Infrastructure Security Agency (CISA) said organizations should begin their transition to post-quantum cryptography now, rather than waiting until cryptographically relevant quantum computers (CRQCs) are available to threat actors.

Why the Post-Quantum Cryptography Shift Cannot Wait

The publication, titled "Preparing for the Post-Quantum Era: A Call to Action," warns that the quantum computing threat is no longer a distant concern. While the exact timeline for CRQC development remains uncertain, the working group said recent technological advances suggest such machines could emerge sooner than expected, putting widely used public-key cryptography mechanisms at risk.

One of the most immediate dangers is a tactic known as "harvest now, decrypt later," where malicious actors intercept and store encrypted data today with the intention of decrypting it once a CRQC becomes available. This poses a serious risk to governmental records, sensitive personal data, and trade or business secrets that require long-term confidentiality.

The advisory also cautions that CRQCs could eventually be used to target authentication mechanisms, allowing bad actors to impersonate trusted entities, forge data, or compromise equipment. Because supply chain vulnerabilities can cascade, a single organization's delay in adopting post-quantum cryptography could expose entire sectors to compromise.

According to the report, organizations that fail to act may also face business consequences beyond security risk, including exclusion from public procurement contracts and loss of competitive advantage.

Five Priorities for the PQC Transition

The G7 Cybersecurity Working Group outlined five priority areas to guide the global shift toward post-quantum cryptography:

1. Raising awareness — Many organizations still view the quantum threat as a distant or purely technical issue. The group called for awareness campaigns, technical guidance, and workforce upskilling to reframe it as an economic and business risk.
2. Developing national strategies — Countries are encouraged to build strategies that ensure an adequate supply of quantum-safe hardware and software while encouraging adoption, integrating the effort into broader digital privacy and security policies.
3. Advancing research and development — Governments should fund research programs and support pilot projects and testbeds to help organizations test and refine their transition to post-quantum cryptography.
4. Building public-private partnerships — Collaboration between government, industry, and academia is seen as key to developing domestic expertise, lowering transition costs, and sharing playbooks and case studies across sectors.
5. Integrating PQC into cybersecurity requirements — The group recommends treating post-quantum cryptography adoption as a natural evolution of cryptographic best practice, and embedding requirements into public procurement to push both vendors and organizations toward quantum-safe systems.

The advisory emphasizes that the shift to post-quantum cryptography cannot be solved by individual organizations in isolation. Instead, it calls for early engagement, coordinated planning, and informed decision-making across public and private sectors worldwide.

Tackling the risks that the impending quantum computing era poses to current cryptographic systems... requires a coordinated global effort to transition to PQC," the report states, adding that public and private organizations must act now to safeguard confidential data, supply chains, and critical systems.

The document was jointly published by cybersecurity authorities from Canada, Germany, Italy, Japan, the United Kingdom, the United States, and France's ANSSI, with participation from the European Commission and support from the EU Agency for Cybersecurity (ENISA).

The U.S. State Department has posted a $10 million reward for Amir Yaryab, a senior Iranian official accused of leading the Islamic Revolutionary Guard Corps Cyber-Electronic Command (IRGC-CEC) Cyber Operations Command and directing multiple hacking groups targeting critical infrastructure across the United States, Europe and the Middle East. According to the Rewards for Justice program, Yaryab allegedly oversees cyber operations conducted by IRGC-CEC-affiliated groups including CyberAv3ngers, Dadeh Afzar Arman (DAA) and Mehrsam Andisheh Saz Nik (MASN). U.S. officials accuse these groups of using malware and conducting cyber and cyber-enabled information operations against civilian infrastructure worldwide.

$10 Million Reward for Amir Yaryab

The $10 million reward for Amir Yaryab seeks information leading to his identification or location. The offer applies to individuals acting at the direction or under the control of a foreign government who participate in malicious cyber activities against U.S. critical infrastructure in violation of the Computer Fraud and Abuse Act. [caption id="attachment_113961" align="aligncenter" width="600"] Image Source: https://rewardsforjustice.net/[/caption] Yaryab is also accused of directing Shahid Hemmat and Shahid Shushtari, two groups linked to cyberattacks against U.S. organizations. The sectors allegedly targeted include defense, news, shipping, travel, energy, financial services and telecommunications. The six Iranian officials named in the advisory are linked to Iran's Islamic Revolutionary Guard Corps and its Cyber-Electronic Command.

Iranian Cyberattacks Target PLCs

The allegations also involve attacks against programmable logic controllers (PLCs), highlighting concerns around Iranian cyberattacks targeting industrial systems rather than focusing only on data theft. U.S. officials said Iranian-linked hackers compromised industrial control systems, specifically targeting the Vision series of PLCs manufactured by Israel-based Unitronics. These devices are used across water and wastewater, energy, food and beverage, manufacturing and healthcare sectors. The attackers exploited default credentials on the devices and left anti-Israel messages. Some of the compromises reportedly rendered the PLCs inoperative. The CyberAv3ngers group, which is linked to the IRGC-CEC, claimed responsibility for attacks against Unitronics Vision PLCs in October 2023. Beginning in November 2023, the group compromised default credentials in PLCs across the United States and left messages on the devices' digital screens.

CyberAv3ngers Attacks Critical Infrastructure

CyberAv3ngers has also claimed responsibility for attacks affecting other infrastructure. In October 2023, the group claimed it had breached ORPAK Systems, a provider of gas station solutions in Israel. The group said it had obtained the company's database and intended to publish it through its Telegram channel. The attack was reported to have disconnected 200 gasoline pumps from the system in the occupied Palestinian territories. In December 2023, CyberAv3ngers also claimed to possess and sell 1TB of data allegedly linked to Israel's electricity infrastructure. The group advertised the dataset for 5 Bitcoin, with an initial 100GB portion also offered at the same price.

U.S. Agencies Warn of PLC Cyberattacks

Concerns over critical infrastructure attacks involving PLCs continued into 2026. A joint advisory issued on April 7 by the FBI, CISA, NSA and other agencies warned that Iran-linked threat actors were actively exploiting internet-facing PLCs. The advisory said several organizations had experienced operational disruptions and financial losses after attackers interfered with industrial processes. The developments come amid broader U.S. actions against Iranian-linked cyber activity. The Justice Department accused Iran-connected hackers of breaching employee email accounts associated with the Department of Labor, the Federal Energy Regulatory Commission and multiple United Nations organizations. The Treasury Department also sanctioned Iranian nationals over cyberattacks targeting critical infrastructure. The State Department's reward offer places Amir Yaryab and the alleged activities of IRGC-CEC-linked groups at the center of the U.S. effort to identify individuals responsible for malicious cyber activity targeting critical infrastructure.

Two Citrix NetScaler vulnerabilities affecting Citrix NetScaler Application Delivery Controller (ADC) and Citrix NetScaler Gateway products have prompted a patching warning for Australian organisations. The Australian Signals Directorate's Australian Cyber Security Centre (ASD's ACSC) has advised organisations using the products to assess their environments and apply available security updates as a priority. Citrix has identified two vulnerabilities affecting NetScaler ADC and NetScaler Gateway, which are critical edge devices used in enterprise networking to securely deliver applications, data and remote access to users.

Citrix NetScaler Vulnerabilities Affect ADC and Gateway

The first flaw, CVE-2026-19489, is a memory overflow vulnerability. According to the alert, exploitation of this vulnerability requires SIP ALG, or Session Initiation Protocol Application Layer Gateway, to be enabled on a Large Scale NAT (LSN) group configuration. The second flaw, CVE-2026-19490, is an authentication bypass vulnerability. The vulnerability requires SAML actions to be enabled and/or the affected product to be configured as a VPN gateway. The conditions required for each vulnerability mean that organisations need to assess their specific Citrix configurations to determine whether affected systems are present in their environments.

Patches Released for Citrix NetScaler products

Citrix released patches for the affected products on August 19, 2026. ASD's ACSC is urging organisations to review the vendor's mitigation guidance, identify vulnerable versions of Citrix products and update affected systems to the latest versions. The advisory places particular emphasis on timely patching because critical edge devices are frequently targeted by threat actors as an entry point into sensitive environments. However, ASD's ACSC said it has no information indicating that a specific Australian industry or sector is currently being targeted in connection with these vulnerabilities.

Organisations Urged to Assess Vulnerable Versions

The mitigation guidance calls on organisations to assess their networks and environments for vulnerable versions of Citrix products and apply patches as soon as practicable. Organisations should also review the mitigation advice provided by Citrix and confirm that affected systems have been updated. Where NetScaler ADC and NetScaler Gateway products are managed by a third party, organisations are advised to contact the relevant managed service provider (MSP) or enterprise IT provider. They should confirm that the products have been patched and are being monitored for suspicious activity. This step is particularly relevant for organisations that do not directly manage their Citrix infrastructure and may rely on external providers for patching and monitoring.

Monitoring Remains Important After Patching

Alongside addressing the Citrix NetScaler vulnerabilities, organisations are advised to monitor affected environments for suspicious activity. The alert recommends notifying ASD's ACSC if suspicious activity is detected. The two vulnerabilities affect different configurations, with CVE-2026-19489 requiring SIP ALG to be enabled on an LSN group configuration, while CVE-2026-19490 requires SAML actions to be enabled and/or the product to be configured as a VPN gateway. For Australian organisations using Citrix NetScaler products, the immediate steps outlined by ASD's ACSC are to identify vulnerable versions, apply the available patches, confirm third-party-managed systems have been addressed and maintain monitoring for suspicious activity.

The CISA and FBI, along with cybersecurity agencies from Australia, Canada, New Zealand and the U.K., have released new guidance on outage communications for service providers dealing with major IT and OT outages. The guide calls for prompt, factual and audience-specific communication during disruptions caused by malicious cyber activity or non-malicious events.

Titled “Communicating Under Pressure: Best Practices for Service Providers,” the guidance says effective communication is critical to limiting operational impact when IT and OT outages affect customers, network defenders, critical infrastructure owners and operators, and the public. It recommends that organizations clearly communicate what is known, what remains unknown and what is still under investigation, while providing frequent updates as circumstances change.

Outage Communications Should Start With Facts

The agencies recommend that service providers establish an outage communications plan before an incident occurs. The plan should define incident thresholds, escalation paths, target audiences and procedures for status pages, customer and partner notices, and regulatory communications. Organizations are also advised to establish cross-functional incident teams involving engineering and operations, communications, legal, risk and compliance, and customer support.

The guidance calls for clearly defined roles, including an incident lead, communications lead and spokesperson. It also recommends parallel workstreams so technical teams can focus on diagnosing and remediating the root cause while communications teams manage external messaging and leadership handles strategy and regulatory requirements.

For organizations responding to cyber incidents, the guidance places particular emphasis on balancing transparency with operational security. If malicious activity is suspected or confirmed, external communications should not compromise investigations, containment efforts or other response activities. Organizations are also advised against making premature conclusions when the root cause remains under investigation.

Service Providers Urged to Tailor Messages

The guidance recommends segmenting communications for technical teams, executives and the public. Audiences can include enterprise IT teams and security operations centers, employees and customers, government partners and regulators, critical infrastructure owners and operators, as well as the media and general public.

During an outage, organizations should lead with a concise summary covering affected systems, user impact, scope and the known cause without speculation. The agencies also advise against vague descriptions such as “service degradation” and recommend messaging that can be understood quickly during high-pressure situations.

Transparency is another central principle. Service providers are advised to state what they know and do not know, use a single source of truth such as a status page, and focus communications on actionable guidance rather than reputation management. Customers should be told what actions they need to take or clearly informed when no action is required.

The guidance also calls for continuous, time-stamped updates that show the incident timeline, actions taken, and recovery milestones. Organizations should maintain a single status page and align external messaging with legal, contractual and sector-specific reporting obligations.

Agencies ultimately frame effective outage communications around five principles: immediate acknowledgement, technical and actionable information, transparency, accountability, and continuous updates. For service providers, the guidance positions communication as an important part of incident response, alongside technical remediation and recovery.

An eight-month international operation targeting West African organized crime groups has resulted in 58 arrests and the identification of 263 suspects across 22 countries, according to INTERPOL. Operation Jackal IV, conducted from November 2025 to June 2026, focused on disrupting criminal networks, tracing illicit funds, identifying high-value targets and supporting arrests and prosecutions. The operation brought together countries across six continents to tackle the growing global threat posed by West African criminal networks, including Black Axe and similar groups. These networks have been linked to a significant share of global cyber-enabled financial fraud, including romance scams, cryptocurrency and investment scams, and business email compromise fraud.

Operation Jackal IV Targets West African Organized Crime Groups

Operation Jackal IV also targeted money laundering activities used to move and conceal criminal proceeds across borders. INTERPOL coordinated cross-border intelligence sharing, analysis and operational support during the operation. It also provided specialized training to strengthen international investigations into financial crime. Tomonobu Kaya, Director of the INTERPOL Financial Crime and Anti-Corruption Centre, said the operation showed the importance of international cooperation in following illicit financial flows and disrupting criminal networks. [caption id="attachment_113806" align="aligncenter" width="600"] Image Source: INTERPOL[/caption]

Major Arrests and Financial Crime Investigations

In Argentina, authorities identified 196 individuals linked to a major Crime-as-a-Service network suspected of providing website domains and money laundering support to West African organized crime groups. The investigation resulted in 17 arrests, with an INTERPOL Operational Support Team assisting with analysis of seized data and identification of suspects and criminal networks. South African authorities raided seven locations in Johannesburg linked to a syndicate involved in romance and investment scams targeting retirees in English-speaking countries. Investigators arrested 39 people, seized USD 2.67 million and blocked 257 bank accounts. In Italy, investigators identified an individual connected to a pan-European money laundering network that used shell companies, remittance services and cash withdrawals. One account processed EUR 845,000, or about USD 736,000, through 560 transactions involving 20 financial instruments. Romanian authorities dismantled a criminal group operating an investment scam through a call centre. The group promoted high returns from stocks and cryptocurrencies, with victims' money transferred to electronic wallets controlled by perpetrators. Authorities estimated that EUR 143 million had been stolen and laundered globally. Eleven people were arrested, while cash, cryptocurrency, six real estate properties and luxury watches were seized.

Sextortion and Crime-as-a-Service Emerge

Beyond individual investigations, the operation highlighted emerging threats involving sextortion and Crime-as-a-Service. INTERPOL identified an increase in West African organized crime groups using sextortion to target minors, including victims as young as 14. In these cases, offenders typically contacted minors through social media, established trust and persuaded them to share explicit images or videos. They then threatened to distribute the material to the victim's contacts unless a ransom was paid. Investigators also found that some criminal syndicates were procuring Crime-as-a-Service from external providers, including through the dark web. These services were used to outsource activities such as money laundering and other operational functions. While several cases from Operation Jackal IV remain under investigation, the preliminary results demonstrate the scale and international reach of the networks targeted during the eight-month operation. The participating countries were Austria, Argentina, Australia, Canada, Côte d'Ivoire, France, Germany, Indonesia, Ireland, Italy, Japan, Malaysia, the Netherlands, Nigeria, Portugal, South Africa, Spain, Sweden, Switzerland, the United Arab Emirates, the United Kingdom and the United States.

The UK Ukraine AI partnership will give Britain access to Ukraine’s Avengers AI Labs, bringing together Ukrainian battlefield experience, operational data and engineering expertise with the UK’s AI ecosystem. The agreement, signed by President Volodymyr Zelenskyy and Prime Minister Andy Burnham in Ukraine, will focus initially on defence and national security. Under the partnership, British innovators and researchers will gain access to data and insights collected across the battlefield. The UK government described Avengers AI Labs as a “goldmine of battlefield data,” offering researchers access to real-world operational information used to train AI models.

How Avengers AI Labs Uses Battlefield Data

The data is collected through thousands of daylight cameras and infrared sensors deployed across the battlefield. The systems capture images and information involving tanks, artillery, air defence systems, infantry and aerial targets, including Shahed drones and reconnaissance UAVs. The data is used to train AI models to recognize and classify battlefield objects. Ukraine’s Defense Ministry has previously said that systems trained using the Avengers Labs platform analyze more than 100,000 drone video feeds each month and help identify about 70% of enemy targets in real time. The UK’s access to the platform is intended to allow British startups, researchers and engineers to work with operational insights and develop technologies based on real-world datasets. The partnership will initially bring together engineers, academics, businesses and military operational expertise from both countries to address national security challenges. The two countries will also explore additional platforms for future collaboration.

UK Ukraine AI Partnership Test New Defence Technology

Several pilot projects involving British startups have already been rolled out as part of the agreement. The companies named are Bristol-based Sintela, Oxford’s Mind Foundry and London’s Skyral. The first technology is due to be deployed at a UK defence site to help protect bases from protestors and hostile actors seeking intelligence. The project combines Ukrainian data with UK technology and turns buried fibre-optic cables into an AI-enabled sensor. The technology could also be used in other critical locations, including airports, prisons, railways and energy plants, according to the information released about the partnership. A second project will examine the development of next-generation low-power AI chips designed for future drones, robotics and autonomous systems. If successful, the technology could support machines designed to operate for longer, respond faster and function in environments where conventional systems face limitations.

AI Sovereignty and Defence Innovation

The agreement forms part of the UK and Ukraine’s 100 Year Partnership and expands cooperation between the two countries in AI and defence technology. The UK will provide access to its universities, researchers, technology companies and AI ecosystem, while Ukraine will provide access to operational experience and datasets generated during the war. Minister for AI Kanishka Narayan described the arrangement as AI sovereignty in practice, focused on developing national capabilities and turning frontline experience into technologies for military and critical infrastructure protection. The partnership also follows the UK government’s announcement that defence firm MBDA can release classified information on UK components for the long-range SCALP missile to establish local assembly lines in Ukraine. The broader agreement is intended to combine Ukrainian battlefield data with British scientific, engineering and technology expertise, with the initial focus remaining on defence, national security and the development of future defence technology.

New Zealand is moving ahead with a social media ban for under-16s, with the Government introducing the Online Safety (Minimum Age and Child Safety Risk Assessment) Bill to Parliament. The proposed law would require high-risk social media platforms to take reasonable steps to prevent children under 16 from accessing their services and would place greater responsibility on technology companies to address risks to children online. Prime Minister Christopher Luxon said the Government was responding to concerns about the amount of time young people spend on social media and their exposure to harmful content, addictive technology and online pressures. “We simply cannot accept the harm being done to a generation of New Zealand children,” says Prime Minister Christopher Luxon. According to the Government, one in three children aged 13 to 17 spends at least five hours a day on social media.

New Zealand Social Media Ban Introduces Age Verification

Under the proposed social media ban, high-risk platforms including Instagram, TikTok, Snapchat and Facebook would be required to take reasonable steps to establish whether users are over 16. The Bill allows for multiple age-checking methods, including existing account information, facial age estimation, digital ID services and formal identification. The proposed framework would also require platforms used by children to regularly assess the risks they pose and report on how those risks are being identified and reduced. The Government said the framework is designed to evolve alongside technology and would also cover emerging services, including AI companion platforms.

Companies Could Face Significant Penalties

The proposed online safety law would introduce significant penalties for platforms that fail to meet their obligations. Companies could face fines of up to 10 percent of their global revenue. An online safety regulator within the Department of Internal Affairs would independently monitor compliance, investigate platforms and enforce the law. Education Minister Erica Stanford said the Bill would place legal obligations on technology platforms rather than children, parents or caregivers. No penalties are proposed for children, their parents or caregivers under the proposed framework. “The Bill places legal obligations on platforms. No penalties are proposed for children, their parents or caregivers,” Ms Stanford says. “It doesn’t just introduce a minimum age requirement. It creates an enduring framework that can evolve alongside technology and holds platforms responsible for understanding and reducing the risks they create for children,” she says. The Government also plans to strengthen education and public awareness so parents, schools and communities have information and tools to help children navigate online services safely.

New Zealand Joins Growing Push for Child Online Safety

The proposed New Zealand social media ban follows similar moves in other countries seeking to place greater responsibility on technology companies. Earlier this year, Spain announced plans to prohibit social media access for children under 16. Prime Minister Pedro Sanchez described the internet as a “digital Wild West” and said stronger protections were needed to shield young people from online harm. France has also moved toward age-based restrictions. In February, French lawmakers approved legislation banning children under 15 from accessing social media platforms, with the measure expected to take effect at the start of the next school year. French President Emmanuel Macron backed the proposal, arguing that children’s development should not be dictated by algorithms designed to maximize engagement.

Government Says Rules Aim to Change Online Behaviour

Luxon acknowledged that the Government does not expect every child to stop using social media but said early evidence from Australia showed minimum age requirements were already changing behaviour. The Government said the child online safety framework is intended to make a meaningful difference while creating longer-term obligations for technology companies. The Online Safety (Minimum Age and Child Safety Risk Assessment) Bill will now be introduced to Parliament, beginning the legislative process for the proposed under-16 social media restrictions.

The U.S. Department of the Treasury has launched Operation Economic Outcast, a whole-of-government campaign aimed at disrupting the economic networks and revenue channels supporting the Iranian regime and the Islamic Revolutionary Guard Corps (IRGC). The initiative expands Iran sanctions across digital assets, technology, gold, aviation and shipping, while targeting nearly 60 entities, individuals and vessels across multiple jurisdictions. Treasury said the campaign follows direction from President Trump and is intended to systematically target financial channels used for oil smuggling, sanctions evasion and other activities linked to Iran.

Operation Economic Outcast Expands Iran Sanctions

Under Operation Economic Outcast, the Office of Foreign Assets Control (OFAC) issued five sectoral sanctions determinations covering digital assets, technology, gold, aviation and shipping. Treasury said the measures increase its ability to sanction foreign persons operating in or providing services to these sectors of the Iranian economy. The department said Iran has increasingly used cryptocurrency for sanctions evasion, while advanced technology has been sought for weapons programs. Gold has also been used to stabilize the rial, while aviation and shipping networks have been linked to the movement of fighters, weapons, sensitive technologies, oil and other assets. The new determinations build on earlier measures covering Iran’s financial, petroleum and petrochemical sectors.

OFAC Sanctions Nearly 60 Iran-Linked Targets

OFAC also sanctioned nearly 60 entities, individuals and vessels across networks associated with nuclear and missile technology procurement, cyber operations and oil revenue generation. The action includes a procurement network spanning the Middle East and East Asia that Treasury said helped Iranian entities obtain sensitive dual-use technology through front companies, financial channels and logistics intermediaries. Treasury also targeted a malicious cyber group directed by Iran’s Ministry of Intelligence and Security (MOIS). The department said members of the group compromised and exfiltrated data from U.S. companies in critical infrastructure sectors, including energy, healthcare, defense, information technology and financial services. The designations also include Iranian cyber actors accused of network compromises and digital asset theft. Treasury said one individual illicitly gained control of a Bitcoin wallet containing more than $30,000 in 2023.

Secondary Sanctions Risk Expands

The campaign also increases secondary sanctions exposure for entities that continue conducting certain business with the Iranian regime. Treasury said countries are being given timelines to address identified Iran-related activity, while entities facilitating money laundering or sanctions evasion could face restrictions involving the U.S. financial system. OFAC also suspended several general licenses that previously authorized certain remittance payments to Iran and Iranian access to parts of the U.S. cultural and academic system.

Iran’s Shadow Fleet and Oil Networks Targeted

A major component of the measures focuses on Iran’s shadow fleet and oil revenue channels. Treasury sanctioned brokers, companies, and vessels involved in transporting Iranian crude oil and petroleum products across multiple jurisdictions. The department identified shipping networks involving the UAE, Hong Kong, China, Singapore, Switzerland, Europe, and other regions. Several UAE-based entities and individuals were designated over alleged roles in facilitating Iranian oil shipments and cryptocurrency payments. OFAC also targeted five vessels identified as blocked property, including SIFRA, G SILVER, QUANTUM HOPE, VOYAGE ELITE and TELA. Treasury said these vessels had transported Iranian LPG, petroleum products or crude oil to markets in Asia. The measures mean that property and interests in property belonging to designated or blocked persons that are in the United States or under the control of U.S. persons are blocked and must be reported to OFAC. Treasury said violations of U.S. sanctions can result in civil or criminal penalties, while certain transactions involving designated persons may also expose foreign financial institutions to secondary sanctions.

Every enterprise wants the upside of AI — faster workflows, sharper decisions, leaner teams. Far fewer have asked the harder question: what happens when the same systems delivering that upside are also quietly rewriting who, or what, has access to the crown jewels? Autonomous agents are now reading contracts, touching customer data, writing production code, and triggering workflows once reserved for trusted employees. The attack surface hasn't just grown — it's changed shape entirely. Mohammad Arif, Head of Information Security at Guild Group, has been on the front line of that shift, helping enterprise leaders separate genuine AI-driven cyber risk from the noise around it. In this interview with The Cyber Express, he makes the case that AI security isn't a niche technical concern to be handed off to a working group — it's a boardroom issue, sitting at the intersection of data protection, vendor risk, identity, and human accountability. His warning is blunt: organisations still treating AI security as tomorrow's problem are already behind, and the cost of catching up only rises from here.

The Cyber Expres: Everyone is talking about AI productivity. From a security leadership perspective, what is the biggest shift AI is creating for enterprises?

From a security leadership perspective, the biggest shift is that AI is changing the enterprise trust model. It is no longer only about protecting systems, networks, and users. Organisations now need to consider what AI can access, what data it can process, what decisions it may influence, and what actions it may trigger. AI can significantly improve productivity, cyber defence, and business decision-making. But the risk increases when adoption moves faster than governance. If AI tools are connected to sensitive data, enterprise workflows, identity systems, APIs, or business processes without clear controls, the risk is no longer just a technology issue. It becomes a data, privacy, operational resilience, and trust issue. So the real shift is this: AI is not just another productivity tool. It is becoming part of the enterprise operating model, and therefore it needs to be governed and secured like any other critical capability.

The Cyber Express: What worries you most about the current pace of AI adoption across organisations?

What worries me most is the gap between the speed of AI adoption and the maturity of AI governance. Many organisations are moving quickly to unlock productivity benefits, but they may not yet have clear rules around approved use cases, sensitive data handling, vendor assurance, retention of information, model outputs, and human accountability. The immediate risk is uncontrolled or "shadow" AI adoption, where employees or teams use public or unapproved AI tools without understanding what data is being entered, how that data may be retained, or whether it could be used to improve external models. This can create confidentiality, privacy, intellectual property, and regulatory risks. The challenge for enterprises is not to slow down innovation unnecessarily. The challenge is to enable AI safely — giving employees approved pathways to use AI, while putting the right controls around data protection, access, monitoring, and risk-based governance.

The Cyber Express: At what point does AI stop being just a productivity tool and become a real enterprise security challenge?

AI becomes a real enterprise security challenge when it moves from simply generating content to being connected to enterprise data, identity, applications, APIs, tools, and workflows. A standalone AI assistant used for drafting general content has a different risk profile from an AI agent that can access customer information, analyse internal documents, write code, trigger workflows, or make recommendations that people rely on. The risk increases further when AI has autonomy, can call external tools, or can operate across multiple systems. This is where agentic AI becomes important. The question is no longer only "what can the model say?" It becomes "what can the AI access, what can it do, and who is accountable for the outcome?" That is the point where traditional security controls need to extend into AI governance, identity, permissions, logging, monitoring, and human oversight.

The Cyber Express: Do you think most organisations are prepared for AI-driven cyber threats, or are many still treating AI security as a future problem?

Preparedness varies significantly. Some organisations are taking AI security seriously and are building governance, security review, data protection, and monitoring into their AI adoption programs. But many are still treating AI security as a future problem, even though AI is already inside the enterprise through productivity tools, SaaS platforms, coding assistants, analytics tools, third-party services, and employee-led experimentation. The issue is that AI adoption is often decentralised. It may start in business teams, technology teams, vendors, or individual users before the organisation has a complete view of the risks. That means security leaders need to shift from reactive control to proactive enablement. AI security readiness should include clear acceptable-use guidance, approved tools, data classification, vendor due diligence, secure development practices, incident response scenarios, awareness training, and monitoring. Without those foundations, organisations may not know where AI is being used, what data is exposed, or where the risk is accumulating.

The Cyber Express: What AI-related security incidents do you realistically expect enterprises to face over the next 12 months?

Over the next 12 months, I expect most enterprises to see AI-related incidents in a few practical areas. The first is AI-assisted social engineering. Phishing, impersonation, business email compromise, and executive fraud will become more convincing because AI allows attackers to generate personalised and credible content at scale. The second is sensitive data leakage into AI tools, which may happen when employees enter confidential business information, customer data, source code, contracts, security information, or internal documents into tools that have not been approved or assessed. The third is insecure AI integration. As teams connect AI to internal knowledge bases, applications, plugins, and workflows, weaknesses such as prompt injection, excessive permissions, poor output validation, or weak monitoring may create new attack paths. The fourth is AI supply-chain risk. Organisations increasingly rely on third-party models, APIs, datasets, plugins, open-source components, and AI-enabled SaaS platforms — each introducing dependencies that need to be assessed and monitored. So the threat is not one single scenario. It is an expanded attack surface across people, data, applications, vendors, and business processes.

The Cyber Express: AI-generated phishing and deepfakes are receiving significant attention. Have these threats become genuinely dangerous, or are they still more hype than reality?

They are genuinely dangerous, but the risk needs to be understood properly. The concern is not only that AI can create fake emails, voices, images, or videos. The bigger issue is that AI reduces the effort required to create believable, personalised, and scalable deception. Traditional phishing often had indicators such as poor grammar, generic wording, or obvious formatting issues. AI reduces those indicators. Attackers can tailor messages to specific roles, business processes, recent events, or organisational language. Deepfakes also create risk where organisations rely heavily on voice, video, or informal executive instructions for approvals or sensitive actions. This does not mean every organisation will face sophisticated deepfake attacks immediately. But it does mean that trust-based processes need to be strengthened. Payment approvals, changes to bank details, privileged access requests, sensitive data transfers, and executive instructions should have strong verification controls that do not rely on one communication channel alone.

The Cyber Express: Do enterprises fully understand the risks of feeding sensitive data, business context, or internal information into frontier AI systems?

Not always. Many organisations understand the general concern, but they may not fully understand the practical ways sensitive information can be exposed through AI usage. The risk is broader than simply entering customer data into a public tool. Employees may enter internal strategies, contracts, security designs, source code, incident information, board papers, commercial terms, or confidential business context. Even if the data is not used for model training, there may still be risks around retention, access, logging, jurisdiction, vendor terms, and downstream use. Enterprises need to apply the same discipline to AI that they apply to other sensitive platforms — understanding what data is allowed, which tools are approved, whether enterprise-grade privacy and security settings are enabled, how data is retained, and whether the vendor's contractual terms align with the organisation's obligations. The practical starting point is data classification. If organisations do not know what data is sensitive, they cannot consistently govern how that data should or should not be used with AI.

The Cyber Express: We have seen supply-chain attacks target software and open-source ecosystems. Could AI models, tools, plugins, agents, datasets, and integrations become the next major supply-chain risk?

Yes, AI supply-chain risk is likely to become a major area of focus. In traditional technology environments, organisations already assess software vendors, cloud providers, managed service providers, open-source libraries, and third-party integrations. AI expands that supply chain. The AI supply chain can include foundation models, fine-tuned models, datasets, model providers, APIs, plugins, orchestration platforms, vector databases, prompt libraries, AI coding tools, agent frameworks, and embedded AI features in SaaS products. A weakness or compromise in any of these areas can affect the confidentiality, integrity, or reliability of AI-enabled systems. A poorly governed dataset could introduce bias or inaccurate outputs. A vulnerable plugin could expose data. A compromised package could affect an AI-enabled application. An over-permissioned AI agent could access more information than it needs. These are not only technical risks; they are third-party, operational, and governance risks. Enterprises should therefore treat AI supply-chain assurance as part of broader cyber risk and vendor risk management, including due diligence, contractual controls, security testing, data protection review, monitoring, and clear accountability between the organisation and its providers.

The Cyber Express: How should security leaders rethink traditional cybersecurity strategies in a world where AI can write code, analyse vulnerabilities, automate tasks, and support attackers as well as defenders?

Security leaders should not abandon traditional cybersecurity principles. Instead, they need to extend them into AI-enabled environments. Identity and access management remains critical, but now we need to consider identities and permissions associated with AI agents, service accounts, APIs, and automated workflows. Data protection remains critical, but now we need to monitor prompts, outputs, embeddings, and knowledge retrieval systems. Secure development remains critical, but now we need to assess AI-generated code, model behaviour, prompt injection risks, and third-party AI components. The same applies to monitoring and incident response. Security teams need visibility into AI usage, unusual activity, sensitive data exposure, misuse of AI tools, and unexpected agent behaviour. Incident response plans also need to consider AI-specific scenarios such as data leakage through AI platforms, compromise of AI integrations, prompt injection, poisoned data, or misuse of AI-generated code. The key point is that AI security should not sit outside the cybersecurity operating model. It should be integrated into governance, architecture, procurement, engineering, monitoring, awareness, and incident response.

The Cyber Express: Are existing cybersecurity teams equipped to handle AI-era threats, or does the industry need new skills and operating models?

Existing cybersecurity skills remain highly relevant, but they need to be expanded. The fundamentals still matter: identity, data protection, secure architecture, vulnerability management, incident response, third-party risk, and governance. However, AI introduces new concepts that security teams need to understand. Security professionals now need AI literacy — how large language models work at a practical level, how AI applications are integrated, how prompts and outputs can be manipulated, how retrieval-augmented generation works, how AI agents interact with tools, and how AI supply chains are structured. The operating model also needs to evolve. AI security cannot be owned by security alone. It requires collaboration between cybersecurity, technology, data, privacy, legal, risk, procurement, HR, and business teams. In many organisations, the most effective model will be a cross-functional AI governance group supported by security-by-design processes. So yes, the industry needs new skills, but not at the expense of existing cybersecurity disciplines. The future is a combination of traditional cyber expertise, AI literacy, risk management, and business enablement.

The Cyber Express: AI vendors promise speed, scale, and automation. Where should organisations avoid over-relying on AI in cybersecurity and business decision-making?

Organisations should avoid over-relying on AI in areas where decisions are high-impact, sensitive, difficult to reverse, or require strong judgement. AI can assist, but accountability should remain human. In cybersecurity, this includes incident severity decisions, containment actions, identity and privileged access approvals, regulatory interpretation, legal assessments, and decisions that could materially affect customers, employees, or critical operations. AI can help summarise information, detect patterns, prioritise alerts, and recommend actions, but those recommendations should be validated by qualified people. There is also a risk of automation bias, where people trust AI outputs because they appear confident or well-structured. That can be dangerous if the output is incomplete, inaccurate, or based on weak context. Organisations need clear rules for where AI can automate, where it can recommend, and where human approval is mandatory. A practical principle is this: the greater the potential impact, the stronger the need for human oversight, auditability, and accountability.

The Cyber Express: If you were advising enterprise leaders today, what are the first three things they should do to prepare for the security impact of frontier AI models?

The first priority is to establish AI governance. Organisations need clear policies on approved use cases, acceptable tools, data handling, human oversight, and accountability. Governance should not be theoretical — it needs to be embedded into procurement, technology delivery, data management, security review, and business processes. The second priority is to protect sensitive data before scaling AI adoption. This means strengthening data classification, access controls, data-loss prevention, retention rules, and monitoring. Enterprises should know what data can be used with AI, under what conditions, and through which approved platforms. The third priority is to integrate AI into the cyber risk management operating model — including third-party risk assessments, threat modelling, secure development, incident response, employee awareness, logging, monitoring, and assurance activities. Security teams should also start preparing for AI-specific risks such as prompt injection, insecure integrations, sensitive data exposure, excessive agency, and AI supply-chain compromise. AI can create significant value, but only if organisations adopt it with discipline. The organisations that treat AI security as a future issue may already be behind.

"AI adoption without governance is not innovation — it is unmanaged risk. The next phase of enterprise security is about controlling what AI can access, what it can do, and who remains accountable." — Mohammad Arif, Head of Information Security, Guild Group

UT San Antonio cyber incident response efforts have prompted the university to delay the start of fall classes by three days, with the semester now scheduled to begin Monday, August 24. The university said the decision will give its teams additional time to restore technology systems and services following attempted unauthorized activity against its academic campus. The activity was detected over the weekend at the edge of the university's network. UT San Antonio said University Technology Solutions (UTS), working with expert partners, took immediate action to contain the activity and protect the campus technology environment.

UT San Antonio Cyber Incident Prompts Class Delay

The university said its investigation has found no evidence that university data was accessed or exfiltrated as a result of the activity. However, UT San Antonio proactively took some systems and services offline while teams evaluated the technology environment and reinforced security safeguards. The move temporarily disrupted access to several services, including connectivity and email. University President Taylor Eighmy said delaying classes until August 24 would allow teams to restore services carefully and ensure the systems students, faculty and staff rely on are operating properly at the beginning of the semester. University operations will continue as scheduled, with faculty and staff maintaining normal work schedules during the restoration process.

University Technology Systems Remain Under Review

The university technology systems affected by the response include connectivity, email and other essential services. UT San Antonio said restoring these services is being handled as a deliberate process while teams assess the environment and strengthen protections. The university's latest update on August 18 said its password reset system was experiencing delays, with further updates expected. The university also said its phone systems were temporarily unavailable, with service expected to return as technology services were restored. UT San Antonio said its technology teams and expert partners continue to work on restoring services and maintaining protections across the academic campus.

UT San Antonio Extends Student Deadlines

The technology disruption has also led UT San Antonio to adjust several deadlines and registration arrangements for Fall 2026. Thepayment deadline was extended to Friday, August 21 at 5 p.m. CT, giving students additional time to submit payments and allowing time for those payments to be processed. Waitlisting was scheduled to end Wednesday, August 19 at 8 a.m. CT. Students who received waitlist notifications to register on August 14 or 15 were automatically being re-added to the waitlist in their original order. Registration remains open, and students who can access their myUTSA Account can continue making changes to their Fall 2026 schedules. The university said One Stop will continue supporting students and providing updates as additional services become available.

Investigation Into Unauthorized Activity Continues

UT San Antonio said the attempted unauthorized activity was identified before reaching core systems. The university's response included taking systems offline as a precaution while its teams evaluated the environment and reinforced safeguards. The university has not disclosed additional details about the nature of the attempted unauthorized activity or identified those responsible. The investigation remains ongoing, while teams continue restoring technology services and assessing the university's environment. UT San Antonio said it will continue communicating directly with students, faculty and staff as services return to normal. Detailed information about the revised academic schedule, student services and other updates is available through the university's Fall 2026 information page. The university now expects to begin fall classes on August 24, three days later than originally scheduled, as it works to restore services and strengthen protections following the incident.

A suspected ARMA cyberattack has targeted Ukraine's Asset Recovery and Management Agency as it prepares to select a manager for assets linked to IDS Ukraine. ARMA said its servers experienced unauthorized interference ahead of the August 22 deadline for applications, prompting an investigation into whether the incident was part of a broader effort to disrupt its operations. The Asset Recovery and Management Agency, known as ARMA, manages assets seized by Ukrainian authorities, including assets linked to sanctioned Russian individuals and alleged collaborators with Moscow.

ARMA Cyberattack Raises Questions Over IDS Ukraine Competition

ARMA said the attack occurred shortly before the August 22 deadline for applications to participate in the competition to select a manager for assets controlled by sanctioned Russian oligarch Mikhail Fridman. The agency said its experts and law enforcement authorities are examining the cyberattack and the events surrounding the IDS Ukraine competition. The Security Service of Ukraine, or SBU, is investigating the recent attack, while a broader National Anti-Corruption Bureau of Ukraine, or NABU, investigation is examining earlier alleged interference. According to ARMA, signs of illegal interference in processes connected to its work have been recorded since spring. These included unauthorized access to the agency's officials' register. ARMA said the combination of cyber incidents, information activity and increased inquiries from some media outlets and members of parliament had raised concerns about a possible coordinated campaign. The agency said investigators must determine whether these events were intended to disrupt its work, create pressure or affect the competition. ARMA has not identified those it believes may have organized or carried out the alleged campaign.

IDS Ukraine Selection Continues Despite Cyberattack

Despite the incident, ARMA said the competition to select the IDS Ukraine asset manager will proceed according to the procedures and timeframe established by law. The deadline for applications is August 22, 2026, with the competition announcement published through Ukraine's Prozorro public procurement system. The agency said it has also started an audit of the financial indicators of seized IDS group assets to support the legality, objectivity and transparency of the transfer process. ARMA said additional information concerning possible unauthorized access to officials' email accounts and official information will be provided to law enforcement authorities for investigation and legal assessment. Acting ARMA Head Yaroslava Maksymenko said the agency would continue the competition despite what it described as information pressure, political interference and attempts to gain unauthorized access to its resources.

Ukraine Investigates Possible Coordinated Interference

ARMA said the latest incident is not being viewed in isolation. The agency pointed to a similar episode earlier this year, when Reuters reported on a cyberattack involving attempts at interference and hacking alongside increased information activity and inquiries. The agency said each event could have an individual explanation, but their timing and combination warranted further investigation. The cyberattack comes as Ukraine continues efforts to prevent sanctioned Russian capital from retaining control over assets seized in the country. ARMA said this includes preventing control through management arrangements, intermediaries or influence groups. Fridman has been sanctioned by Ukraine and several Western governments since Russia's invasion. ARMA said the final responsibility for determining the organizers, customers and perpetrators of the attack rests with the ongoing investigations. The agency said it will continue the IDS Ukraine competition and act within the law while law enforcement agencies examine the reported cyber incidents and possible attempts to interfere with its activities.

A DGFiP cyberattack has exposed sensitive tax and cadastral information after attackers allegedly used stolen credentials to access systems belonging to France's Directorate General of Public Finances. The French Public Finances Directorate said investigations found that data linked to 678,000 individuals and professionals had been consulted and extracted during intrusions in June and July 2026. The DGFiP cyberattack incidents were identified after a malicious actor claimed illegitimate access to the French tax authority's information system on August 12 and 13. DGFiP said the intrusions involved the usurpation of identifiers belonging to a DGFiP agent and an authorized third party.

DGFiP Cyberattack Exposed Taxpayer Information

After detecting the intrusions, DGFiP immediately suspended access to the accounts involved. Initial access controls did not identify data theft, which the authority attributed to the sophistication of the attack. A subsequent investigation established that the compromised access points had been used to consult and extract information concerning 678,000 individuals and professionals. The exposed information included reference tax income, family quotient and withholding tax rate for individuals. For businesses, the accessed information included company names and SIREN numbers. Cadastral data, including addresses and property sizes, was also accessed. DGFiP said online accounts belonging to individual and professional users were not compromised, and user IDs and passwords were not affected. The authority notified France's data protection regulator, CNIL, after identifying the data breaches.

Cadastral Data Leak Claim Targets DGFiP

Separately, a hacker using the alias ZeroBytes claimed an attack against DGFiP's Professional Cadastral Data Server (SPDC). According to the claim cited by FrenchBreaches, the alleged extraction contains 252,149 lines of data representing 2,041,778 people, with multiple holders potentially associated with the same property plot. The claimed dataset reportedly includes names, surnames, sex, dates and places of birth, addresses, land identifiers, cadastral sections and parcel numbers, as well as information about rights held on properties. The claim would therefore link individuals to personal information and real estate assets. However, the figures and technical details in this second claim remain allegations by the cybercriminal. The claim that the system could contain information relating to approximately 20 million citizens is also an estimate made by ZeroBytes and does not establish that this number of people was affected.

Investigation Into French Tax Authority Attack Continues

DGFiP said additional security measures were implemented after investigators uncovered new information. These included preventative shutdowns of access to sensitive information systems. Investigations remain underway to determine the precise nature and volume of data extracted and the number of users affected. DGFiP teams are working with France's economic and financial ministries, the High Official for Defence and Security and the National Agency for Information Systems Security, ANSSI. The authority said it will contact affected individuals and professionals directly from the following week by email or letter. Those notifications will identify the information that may have been accessed or extracted and outline any precautionary measures where applicable. DGFiP also said it will file a complaint and provide further information as the investigation progresses. The separate cadastral data breach claim remains subject to confirmation, including the alleged number of affected people, duration of access, methods used to bypass authentication, the full scope of extracted information and whether access remained active when the claim was published. The confirmed DGFiP investigation and the separate ZeroBytes claim therefore present different sets of figures and allegations, with the full scope of the incidents still being determined.

Recent AI security incidents involving model evaluations have raised questions about how securely frontier AI models are tested before deployment. Cybersecurity-focused AI company Irregular said an investigation into a publicly reported incident found that models in a small number of evaluation runs gained unintended internet access and carried out offensive security actions against real-world targets. Irregular said the incidents traced back to a single evaluation scenario and were not separate events. The company said the issue was resolved before the initial public disclosure, affected parties were notified, and there are currently no active issues. Its broader audit remains ongoing. Irregular said the AI security incidents occurred during AI model evaluations designed to test whether models could autonomously execute multi-stage cyber scenarios under realistic conditions. The evaluations simulate real-world attack patterns and provide models with objectives such as reconnaissance, identifying private keys, gathering information, extracting data and attempting to avoid detection. According to Irregular, controlled internet access can be important to maintain realism because real-world attackers rely on the internet. However, the company found that internet access controls allowed some models to move beyond the intended simulated environment. In the evaluation at the center of the investigation, engineers selected a fictional company name that unknowingly matched a real domain. While models operated inside the simulation in most runs, a small number attempted to access the real domain despite receiving no instruction to do so. Irregular said models subsequently exploited vulnerabilities, extracted credentials and obtained access to a production database. In another instance, a model moved to a website with a similar name and encountered publicly posted credentials. The company said there is no evidence that a customer's systems were breached or that customer data was leaked.

Evaluation Environments Face Monitoring Challenges

Irregular said most of the problems identified during its investigation were connected to internet access controls. The company is introducing additional protocols for evaluation setup and strengthening its approach to containment. The incident also highlighted challenges around model monitoring. Irregular said advanced simulations can generate extremely large volumes of traffic, while legitimate offensive actions performed by models can resemble malicious activity. The company said incidents occurred in fewer than one in 10,000 advanced simulations and often appeared hundreds of turns into an evaluation. Irregular plans to expand manual review of model behavior, improve monitoring capabilities and establish a dedicated internal team focused on security, containment and model control. The company also said evaluation environments need continuous review because fictional names used in scenarios can later overlap with newly created websites or real domains.

Security Experts Question Irregular's Response

The disclosure has drawn criticism from cybersecurity professionals. Zack Korman, chief executive of cybersecurity-focused AI company Embroidery, called the post “such an embarrassing post-mortem” on the OpenAI/Anthropic security incidents and described it as “full of excuses.” Justin Elze, chief technology officer at TrustedSec, questioned why the monitoring challenge had not been addressed earlier, saying the issue appeared closely connected to the purpose of the testing service. Woodward also criticized the disclosure, arguing that the absence of dates, named owners for corrective measures and independently verifiable criteria limited its usefulness to researchers and security professionals. Another cybersecurity commentator, BlackRoomSec, challenged Irregular's assessment of existing monitoring tools, arguing that security teams routinely tune monitoring systems to reduce noise and filter false positives. Irregular said it is continuing its investigation and will share additional findings where relevant. The company also plans to publish an open whitepaper covering pre-deployment evaluations, including proposed best practices for internet access during testing. [caption id="attachment_113686" align="aligncenter" width="555"] Image Source: X[/caption] The discussion adds to wider scrutiny of how cyber evaluations are contained, monitored and disclosed as frontier AI models become increasingly capable. Irregular said the lessons from the incident will be used to develop stronger evaluation environments, monitoring capabilities, containment controls and response procedures.

The FBI is warning the public about sexual exploitation actors illegally accessing social media and personal accounts to steal explicit images and videos from adult and underage victims. The stolen material, also known as non-consensual intimate images (NCII), is being posted or sold on criminal marketplaces, often without the victim's knowledge.

According to the FBI, these actors use social engineering and cyber intrusion tactics to target specific individuals or general targets of opportunity. After gaining access to accounts, they steal explicit content and share it through community forums or illicit marketplaces.

The FBI said personally identifiable information, including a victim's name, date of birth, email address, phone number and social media username, is often posted alongside the stolen material. This can expose victims to continued harassment and re-victimization.

How Sexual Exploitation Actors Access Accounts

The FBI has identified several methods used by sexual exploitation actors to gain access to victims' accounts.

Password and PIN Targeting

In password/PIN targeting, actors use high-volume password and PIN attempts against social media and personal accounts. The information used in these attempts can come from data leak sites, social media and open-source information.

When victims are known to the actors, curated lists may include personal details such as names, date of birth or variations of those details.

Social Media Customer Service Impersonation

Another tactic involves social media customer service impersonation through text messages. Victims may receive messages claiming their account is being disabled or locked unless they provide a verification code.

The actor then requests a password reset, causing a code to be sent to the victim. If the victim shares the code, the actor can reset the password and access the account.

Phishing Emails

The FBI also warns about phishing campaigns using look-alike domains and email accounts designed to appear as social media customer support.

These messages may claim there has been a new login and contain an embedded link asking the victim to change their password. Clicking the malicious link can give the actor access to the account.

Stolen Content Can Lead to Further Attacks

Once explicit content is stolen, sexual exploitation actors may post or sell it while including personal information about the victim. The FBI said victims can subsequently face harassment, sextortion, stalking or other targeted attacks.

The actors may also advertise stolen content through a victim's own social media page, increasing the potential for further exposure.

FBI Shares Steps to Protect Accounts

The FBI advises people to avoid storing sensitive images or videos on social media platforms or other internet-accessible sites.

It recommends using unique, complex passphrases and PINs along with multi-factor authentication (MFA). Password information directly associated with a person's identity, including names or birthdays, should be avoided.

Users should also be cautious with links received through emails and text messages. The FBI recommends going directly to the relevant website to address account concerns and checking URLs before clicking.

Unrequested temporary passwords, PIN resets or access codes should also be treated with caution. The FBI advises users not to share login information, even when someone claims to represent a platform or service.

People who believe their explicit content was stolen or leaked can provide information through the FBI's NCII reporting site. The FBI also advises the public to continue reporting fraud, scams and cyber threats to the Internet Crime Complaint Center or a local FBI Field Office.

Gunra ransomware has expanded its operations through a structured ransomware-as-a-service (RaaS) affiliate program, prompting the FBI, CISA and other agencies to issue a joint advisory warning organizations about the threat. The Gunra ransomware variant uses a double-extortion model, encrypting victim data while threatening to publish stolen information on a dedicated leak site if ransom demands are not met. The FBI first observed Gunra in April 2025 as a double-extortion ransomware variant derived from leaked Conti ransomware source code.

Gunra Ransomware Shifts to Affiliate Model

By early 2026, the group had expanded through a formal ransomware-as-a-service affiliate program advertised on dark web forums. The program provides affiliates with a management panel, configurable ransomware builder, cross-platform locker payloads and affiliate documentation. The FBI also observed Gunra operating under new branding aliases, including Golden Community, while recruiting penetration testers and ethical hackers as initial access brokers. Gunra initially focused on Windows environments before introducing a Linux variant and moving toward broader cross-platform targeting. Victims observed on the group’s dedicated leak site include organizations across the Americas, Europe, the Middle East, Africa and the Asia-Pacific. Targeted sectors include healthcare and public health, financial services and insurance, critical manufacturing, transportation, government services, utilities, academia, media and communications, retail, and professional and nonprofit services.

VPN Vulnerabilities Used for Initial Access

According to the advisory, Gunra actors primarily gained initial access by exploiting known vulnerabilities in internet-facing devices, including firewall and VPN gateways. The FBI observed exploitation of CVE-2024-55591 and CVE-2025-24472, authentication bypass vulnerabilities affecting specific FortiOS and FortiProxy versions. The Republic of Korea’s National Police Agency also observed Gunra actors exploiting credential exposure and SSH access control weaknesses in internet-facing VPN gateways to obtain unauthorized remote access. After gaining access, attackers used tools including Impacket utilities to move laterally through victim networks using SMB. In one case, actors compromised an SSL-VPN appliance using default credentials where account lockout controls were absent. They later used stolen session information to access internal virtual desktop infrastructure and move through systems including Active Directory servers and IT personnel workstations.

Data Theft Precedes Encryption

The double-extortion ransomware operation involves stealing sensitive information before encrypting systems. The FBI observed Gunra actors collecting business-critical documents, databases, personally identifiable information, and internal email communications. In at least one case, the actors used a malicious executable called main.exe to exfiltrate data from Microsoft OneDrive and SharePoint. Compressed archives containing sensitive information were also transferred to the Mega file-sharing service, with the volume of exfiltrated data reaching tens of terabytes. For encryption, Gunra uses ChaCha20 and RSA-4096 algorithms and has been observed using the .ENCRT extension for encrypted files. A documented sample from July 2025 used the .CRYPT extension. The ransomware also uses Windows Management Instrumentation to delete volume shadow copies before encryption, while one victim had backup and archived data deleted from both primary and disaster recovery infrastructure.

Agencies Urge Patching and Network Segmentation

The authoring agencies recommend that organizations prioritize patching known exploited vulnerabilities in internet-facing systems, including VPN gateways and RDP-exposed infrastructure. They also advise implementing and testing offline, immutable backups stored in physically separate and segmented locations. Network segmentation is another key recommendation, intended to restrict lateral movement and limit the spread of ransomware between systems. The agencies also recommend reviewing domain controllers, servers, workstations and Active Directory environments for unrecognized accounts, auditing administrative privileges, requiring MFA where possible and testing security controls against the Gunra techniques mapped to the MITRE ATT&CK framework. The joint advisory was published August 10, 2026, as part of the ongoing #StopRansomware initiative.

New Zealand has announced a new round of sanctions against Russia, targeting 33 individuals and entities accused of supporting Moscow’s war against Ukraine. The latest New Zealand sanctions against Russia place particular focus on cyber actors, individuals linked to the forced relocation and re-education of Ukrainian children, and entities supporting Russia’s military-industrial complex. Foreign Minister Winston Peters said the package also targets individuals involved in creating and spreading anti-Ukraine propaganda, as well as actors from the Democratic People’s Republic of Korea (DPRK) and Iran providing support to Moscow. “Children should never be used as instruments of war,” Peters said, expressing concern over efforts to abduct and re-educate Ukrainian children through state-directed programmes.

New Zealand Sanctions Target Russian Cyber Actors

The latest Russia sanctions include several individuals previously accused by the United States and other Western governments of malicious cyber activity. Among them are Yuliya Pankratova and Denis Degtyarenko, members of the pro-Russian hacktivist group Cyber Army of Russia Reborn (CARR). The U.S. Treasury sanctioned both individuals in 2024 over alleged cyber operations targeting U.S. critical infrastructure. U.S. officials identified Pankratova, who uses the alias “YUliYA,” as the group’s leader, while Degtyarenko, known as “Dena,” was described as one of its primary hackers. American officials alleged that Degtyarenko was responsible for compromising an industrial control system at a U.S. energy company and had developed training materials for compromising supervisory control and data acquisition (SCADA) systems. Pankratova has also allegedly been associated with Z-Pentest, another pro-Russian hacking group accused of targeting critical infrastructure. New Zealand has also sanctioned Aleksandr Volosovik, known online as “Yalishanda.” U.S. prosecutors have accused him of helping operate Media Land, a Russian bulletproof hosting provider allegedly used by cybercriminals to target organizations including hospitals, schools and banks. In July, the U.S. Justice Department unsealed charges against Volosovik and two other Russian nationals, alleging activities that caused more than $62 million in losses to victims in the United States and other countries.

GRU-linked Official Among Sanctioned Individuals

Another individual included in the latest New Zealand sanctions against Russia is Andrey Averyanov, a senior Russian military intelligence officer who previously commanded GRU Unit 29155. Western governments have linked the unit to cyberattacks, sabotage and other covert operations. Its cyber division has been accused of targeting governments, defense organizations, think tanks and other entities in Ukraine and NATO countries. New Zealand had previously sanctioned members of the unit over alleged malicious cyber activity targeting Ukraine and other countries.

Russia Propaganda and Technology Entities Targeted

The new sanctions also target Russia’s Internet Development Institute (IRI), a Kremlin-backed organization that finances digital media and content promoting Russian state narratives. IRI director Alexey Goreslavsky has also been designated. The British government has previously said the institute was established by Russia’s presidential administration and received hundreds of millions of dollars in government funding. Its projects have included films and video games promoting narratives associated with the Kremlin. Russian information technology company LANIT has also been added to the sanctions list. The company has provided services to Russia’s Defense Ministry and sanctioned defense-industry companies, including state-owned conglomerate Rostec. LANIT had previously been sanctioned by the United States, Canada and Ukraine.

New Zealand Reaches 36th Russia Sanctions Round

The latest package represents New Zealand’s 36th round of Russia sanctions since the Russia Sanctions Act came into force in March 2022. New Zealand has now imposed sanctions on more than 2,000 Russian individuals, entities and vessels, alongside trade restrictions. The measures generally include asset freezes and travel bans and prohibit New Zealanders from making funds or other assets available to designated individuals and entities. Peters said cyber activity can have real-world consequences, noting that cyber actors are increasingly being used to gather intelligence, enable sanctions evasion and disrupt those opposing Russia’s aggression.

A Ransom Cartel ransomware leader has been sentenced to 16 years in prison after being convicted of conspiracy to commit offenses against the United States, conspiracy to commit wire fraud, and aggravated identity theft, according to the U.S. Department of Justice. Maksim Silnikau, a 40-year-old Belarusian national, was identified in court documents as the creator and administrator of the Ransom Cartel ransomware strain, which was developed in 2021. Silnikau had been active on Russian-speaking cybercrime forums since at least 2005 and was also a member of the cybercrime website Direct Connection between 2011 and 2016.

Ransom Cartel Ransomware Operation

Beginning in May 2021, Silnikau developed the ransomware scheme and recruited participants through cybercrime forums. He distributed information and tools to participants, including stolen credentials linked to compromised computers and tools designed to encrypt or lock those systems. Silnikau also maintained a hidden website used by himself and his co-conspirators to monitor and control ransomware operations. According to the court documents, the website provided a platform for the group to communicate with one another, interact with victims, send and negotiate ransom demands, and manage the distribution of funds among the conspirators. The operation targeted companies between 2021 and 2023. During that period, Ransom Cartel ransomware conspirators carried out attacks against at least 18 companies around the world, including organizations based in California, New York, Nebraska, and other countries outside the United States.

Ransomware Attacks Targeted Companies Worldwide

The ransomware attacks involved data theft and monetary demands. The attackers sought payment in exchange for providing keys to unlock stolen data or for promises not to publish the information taken from victims. The operation’s growth was disrupted following Silnikau’s arrest in July 2023. He was later extradited from Poland to face prosecution in the Eastern District of Virginia and the District of New Jersey. The sentencing announcement was made by Theophani K. Stamos, First Assistant U.S. Attorney for the Eastern District of Virginia; Acting Special Agent in Charge Andrew Forrest of the U.S. Secret Service Criminal Investigative Division; Chris Ormerod, Special Agent in Charge of the FBI Kansas City Field Office; and Craig L. Tremaroli, Special Agent in Charge of the FBI Albany Field Office.

Maksim Silnikau Sentenced to 16 Years

The Justice Department’s Office of International Affairs provided substantial assistance with Silnikau’s extradition and the collection of evidence. The U.S. Attorney’s Office for the District of New Jersey and the Computer Crime and Intellectual Property section also assisted with the case. Assistant U.S. Attorney Jonathan S. Keim and former Assistant U.S. Attorney Zoe Bedell prosecuted the case. The 16 years in prison sentence follows the disruption of an international ransomware operation that targeted at least 18 companies during its active period. Silnikau’s role, according to court documents, extended across the development and administration of the ransomware strain, recruitment of participants, provision of attack tools, victim communications, and management of funds generated through the operation.

Levi Strauss cyberattack has exposed certain corporate information after an unauthorized third party gained access to company files through compromised employee computers, according to a filing with the U.S. Securities and Exchange Commission. Levi Strauss & Co. said it recently detected a cybersecurity incident involving unauthorized access to three company-issued computers. The company said the access was enabled through social engineering techniques, allowing the attackers to reach company files. According to the filing, the company initiated its response protocols after detecting the incident and implemented containment measures. It also launched an investigation that remains ongoing and engaged third-party cybersecurity experts to assist with the response.

Levi Strauss Cyberattack Investigation Remains Ongoing

Based on preliminary findings, Levi Strauss said it believes certain corporate information was accessed and exfiltrated during the incident. However, the company said its rapid response efforts successfully contained and terminated the unauthorized access. The company also stated that no consumer data had been impacted as of the date of the filing. Levi Strauss said the incident has not interrupted its business operations and that, based on the information currently available, it does not believe the incident has had or is reasonably likely to have a material impact on its business strategy, operations, financial condition, or results of operations. The company said it has provided and will provide notifications to affected parties and applicable regulators as appropriate and in accordance with applicable law. Levi Strauss & Co., headquartered in San Francisco, is known for its Levi's denim brand. The company reported net revenues of $6.3 billion for 2025, up 4% compared with fiscal year 2024 and 7% on an organic basis. The company designs and markets jeans, casual wear and related accessories for men, women and children under the Levi's, Levi Strauss Signature, and Beyond Yoga brands. Its products are sold in approximately 120 countries through chain retailers, department stores, online sites, and approximately 3,300 retail stores and shop-in-shops.

Retail Cybersecurity Incidents Continue

The Levi Strauss cyberattack comes as several major retailers have reported cybersecurity incidents involving their own systems or third-party service providers. In the first week of August 2026, the De Bijenkorf cyberattack affected the Dutch luxury department store chain after an incident involving one of its external logistics partners. The disruption affected order processing, deliveries, returns, and refunds, while the company said there was no evidence that its own infrastructure had been compromised. In October 2025, Spanish fashion retailer Mango confirmed a Mango data breach after an external marketing service provider experienced unauthorized access to limited customer information. Mango said its corporate systems were not compromised and that financial or login details remained secure. The exposed information included customers’ first names, countries, postal codes, email addresses, and phone numbers. The company said last names, banking information, credit card details, and passwords were not affected. Retailers also faced law enforcement action following a series of attacks. In July 2025, the UK’s National Crime Agency arrested four people suspected of orchestrating cyberattacks against Marks & Spencer, Co-op, and Harrods. The suspects were detained in the West Midlands and London and faced charges under the Computer Misuse Act, blackmail, money laundering, and involvement in an organized crime group. Meanwhile, in May 2025, Victoria’s Secret took down its U.S. website and some in-store services following what it described as a Victoria’s Secret security incident. The company said the precautionary shutdown was intended to address the incident while its team worked to restore operations. Its Victoria’s Secret and PINK stores remained open. The latest incident involving Levi Strauss adds another case to a growing series of cybersecurity incidents affecting major retailers, with the company continuing its investigation into the access and information involved.

OpenAI models accessed the public internet during separate third-party cyber evaluations conducted by independent testing partners, prompting the company to review how high-risk AI testing is managed. OpenAI said the incidents occurred under specialized testing configurations with reduced safeguards and did not reflect how its models operate in public deployments. The company added that the events were unrelated to the previously disclosed Hugging Face security incident.

The incidents involved evaluations conducted by UK AISI and cybersecurity testing partner Irregular, where testing conditions or environment configurations enabled models to interact with systems beyond the intended evaluation boundaries.

OpenAI Models Prompt Review of Third-Party AI Testing

OpenAI said independent cybersecurity evaluations are essential for understanding model capabilities before deployment. Some evaluations intentionally reduce safeguards or enable additional capabilities to measure how models perform under conditions that resemble real-world cyber operations.

According to the company, the latest incidents highlighted the need to strengthen the security controls surrounding independent testing environments as AI models become more capable.

OpenAI said it will review how it manages third-party cyber evaluations, including how higher-risk testing is identified, when internet access or reduced safeguards should be permitted, how testing environments should be isolated, and how incident reporting and monitoring procedures should be handled.

The company also plans to work with national AI institutes, independent evaluators, and other AI developers to establish stronger industry practices for high-risk AI evaluations.

UK AISI Identified Two Unsanctioned Actions by GPT 5.6 Sol

On August 3, UK AISI, the UK's AI Security Institute, informed OpenAI that during a cyber evaluation that began on July 25, models from OpenAI and another AI lab performed activities beyond the intended testing scope.

The evaluation used simulated cyber ranges where models acted as cybersecurity experts in a Capture-the-Flag exercise. Internet access was intentionally enabled so agents could download tools similar to those available to real attackers, while cyber classifiers were disabled to measure underlying model capability.

Among 19 recorded events, two involved GPT 5.6 Sol.

According to OpenAI, the model reused a publicly accessible GitHub token left behind by another lab's agent to determine whether systems inside the simulated environment were communicating with GitHub or update services. It also attempted account recovery and request-limit workarounds while registering accounts with external DNS and tunneling providers.

In a separate action, GPT 5.6 Sol used a public tunneling service to expose a locally hosted DNS server from the evaluation environment to the public internet. The server contained exploit payloads intended for a known software vulnerability within the simulated cyber range.

OpenAI said the setup did not function as intended, no evidence indicated that any real internet resolver queried the server, and the infrastructure was removed when the evaluation concluded.

UK AISI considered both actions outside the authorized testing boundary because they involved external services that were not required to complete the exercise.

Security monitoring detected unusual data transfers on July 28. UK AISI halted the evaluation, isolated the affected systems, and contained the activity within approximately one hour.

Irregular Finds Misconfigured Testing Environment

A second incident was reported on July 29 by Irregular, another organization conducting third-party cyber evaluations for OpenAI.

The testing environment was intended to remain isolated from the internet during another Capture-the-Flag exercise. However, a configuration error unintentionally allowed the models to access external websites.

During one evaluation, the fictional domain used in the exercise matched an actual website. Because internet connectivity was available, the model interacted with the live site, believing it was part of the simulated challenge.

Irregular said the incident did not involve a sandbox escape or zero-day exploit. Instead, the model exploited a basic security vulnerability on the website after gaining unintended internet access.

The investigation also found that the model used credentials associated with the same website. According to Irregular, no impacts beyond that site's own data have been identified, although its audit remains ongoing.

The organization has paused the evaluations, completed remediation, notified affected third parties, and implemented additional safeguards in its testing environment.

OpenAI said it will continue working with both UK AISI and Irregular to improve evaluation practices while ensuring independent cybersecurity testing remains rigorous as AI capabilities continue to advance.