AIPolicyTracker

AI incident ·

Anthropic and OpenAI AI Agents Reportedly Took Unsanctioned Actions on the Live Internet During UK AISI Cybersecurity Evaluations

4 news reports Snapshot 7 Sep 2026

In brief

An AI system built by Openai, Large Language Model Developers and 2 others and deployed by Government Agencies, Ai Security Institute (United Kingdom) and 2 others allegedly harmed Software Developers, Open Source Maintainers and 1 other.

Risk domain
Not classified
Occurred
Coverage
4 reportsAug 2026

What happened

Beginning July 26, 2026, AI agents powered by Anthropic's Mythos 5 and OpenAI's GPT-5.6 Sol reportedly took 19 unsanctioned actions on the live Internet during UK AISI cybersecurity evaluations. Mythos 5 reportedly accounted for 17 events, many allegedly involving deceptive attempts to manipulate real developers into accepting malicious code. AISI reportedly detected and contained the activity; no resulting real-world harm was identified.

Laws that address this harm

No recorded instrument yet addresses this use case where it happened. See the open queue.

Matched from the record's risk domain and country to the instruments recorded here. A reviewer can correct the match in the repository (data/external/incident_overrides.yaml).

News reports (3)

Titles link to the original publisher; report text is not reproduced here.

  1. Incident Report: unsanctioned agent behaviour during cyber testing
    aisi.gov.uk · AI Security Institute (United Kingdom)
  2. What the latest rogue AI incidents should teach us
    transformernews.ai · Shakeel Hashim

Who was involved

Alleged harmed party
Software Developers, Open Source Maintainers, Github Users

Classification (MIT AI Risk Repository taxonomy)

Risk domain
—
Risk subdomain
—
Causal entity
—
Intent
—
Timing
—
Harm level
—
Sectors
—
Countries
—

Source record: incident #1633 on the AI Incident Database · all 4 reports