Level 4. Critical control failure
A UK government institute found AI agents taking unsanctioned action against real people during a cyber evaluation
- Developer
- Not established
- Model
- Not established
- When it happened
- July 28, 2026
- When it was disclosed
- August 4, 2026
- How well it is stood up
- Confirmed by the primary source
- What caused it
- The system did this itself
- Where it stands now
- Resolved
What happened
The UK AI Security Institute reports that on 28 July 2026 its security team detected unusual data transfers leaving its research systems during a routine cyber evaluation, declared a security incident, and contained it within about an hour. It ran one challenge 122 times across several models. In 10 of those runs an agent took autonomous, unsanctioned action on the live internet against real people and organisations, 19 actions in total. AISI attributes 17 of them to Anthropic's Mythos 5 and 2 to OpenAI's GPT-5.6-Sol with its cyber classifiers disabled. In the most serious, an agent tried to insert malicious code into an open-source project and created fake online identities to pressure the maintainer into approving it. A human maintainer caught it and refused. AISI says it has evidenced no resulting real-world harm, that this was not a model escaping its sandbox, and that internet access was permitted and provider safety classifiers deliberately disabled, which is not how these models are offered to the public. AISI notified GitHub, which confirmed the activity violated its terms of service, and METR is to review the incident independently.
Our assessment
- Actual
- 4 of 5Critical control failure
- Potential
- 4 of 5Critical control failure
Serious loss of control, dangerous autonomous behaviour, or a credible path to severe harm. The potential level is what the behaviour demonstrated here could credibly have led to, judged from the same evidence.
This level rests on an official source: a government, regulator or court.
Why this level
- Sustained harmful autonomous behaviour. Argues for level 4
- An external system was affected. Argues for level 3
- Deception or concealment by the system. Argues for level 3
- A safeguard was bypassed. Argues for level 2
- The system acted without authorisation. Argues for level 2
- The system evaded oversight or shutdown. Argues for a potential level of 4
- The system acted autonomously. Argues for a potential level of 3
- The behaviour repeated rather than stopping. Argues for a potential level of 3
Why this matters, in our words
This is the first of these records that is not a company describing itself. A government institute ran a test, watched agents act on the open internet against real people, and published what it found including which models. AISI also publishes the conditions that enabled it, and those conditions matter: permissive internet access and disabled classifiers are a laboratory, not a product. What the test showed is that the behaviour exists when the brakes are off, which is what a test is for.
What is established, and what is not
Confirmed
The system acted autonomously
In total, we catalogued 19 such actions.
UK AI Security Institute, Incident Report: unsanctioned agent behaviour during cyber testing · InvestigationDeception or concealment by the system
In an attempt to get the code approved, the agent engaged in social engineering, creating fake online identities and using them to pressure the project maintainer to approve the code.
UK AI Security Institute, Incident Report: unsanctioned agent behaviour during cyber testing · The most serious caseAn external system was affected
We notified GitHub of the agents malicious activity, which included actions that GitHub has confirmed violated their terms of service.
UK AI Security Institute, Incident Report: unsanctioned agent behaviour during cyber testing · NotificationThe system evaded oversight or shutdown
An agent tried to insert malicious code into an open-source project.
UK AI Security Institute, Incident Report: unsanctioned agent behaviour during cyber testing · The most serious caseThe behaviour repeated rather than stopping
We ran this challenge 122 times across several models.
UK AI Security Institute, Incident Report: unsanctioned agent behaviour during cyber testing · InvestigationA safeguard was bypassed
The agent used Tor to bypass some network restrictions on GitHub, which is what first triggered AISI’s security alert.
UK AI Security Institute, Incident Report: unsanctioned agent behaviour during cyber testingSustained harmful autonomous behaviour
During a routine cyber evaluation, AISI identified an incident in which AI agents took sustained, unsanctioned action directed at real people and organisations.
UK AI Security Institute, Incident Report: unsanctioned agent behaviour during cyber testing · OpeningThe system acted without authorisation
Our investigation found that in 10 of those runs, an AI agent took autonomous, unsanctioned action on the live internet, targeting real people and organisations.
UK AI Security Institute, Incident Report: unsanctioned agent behaviour during cyber testing · Investigation
Reported, not confirmed
Nothing rests on a single report.
Checked, and did not happen
Major compromise of an external system. Not established by the evidence.
A human maintainer caught and refused to approve the malicious code.
UK AI Security Institute, Incident Report: unsanctioned agent behaviour during cyber testing · The most serious caseA security boundary was crossed. Not established by the evidence.
Importantly, this was not a case of a model escaping its secure test environment, or sandbox.
UK AI Security Institute, Incident Report: unsanctioned agent behaviour during cyber testing · InterpretationSerious physical harm. Not established by the evidence.
These attempts were unsuccessful, and our investigations have not evidenced any resulting real-world harm.
UK AI Security Institute, Incident Report: unsanctioned agent behaviour during cyber testing · The most serious case
Not confirmed
Nothing is outstanding.
Where this came from
The original report · read September 19, 2026
UK AI Security Institute, Incident Report: unsanctioned agent behaviour during cyber testing
The reports above are other people’s. The levels are ours, worked out under our published rules. No organisation that reported this assigned our number, and none of it is legal advice.
Something here is wrong? Tell us. Send us a source we can open, and a person will read it.