Latest
  1. Johns Hopkins Bayview tells 3,182 patients an employee put their health data into a public AI tool
  2. China’s AI compute access expected to double in 2027, with US-designed chips about 65 per cent
  3. UK commits £100m to national AI Biosecurity Lab
  4. Moonshot AI releases Kimi K3.1 Pro with 1-million-token context window
  5. Residents in Newport and Cardiff oppose hyperscale datacentres near homes
  6. Study: Grok 4.7 picked Republicans in 92% of forced midterm answers, GPT-6.1 Sol picked Democrats in 97%
  7. Attackers exploiting the PHP flaws hand the break-in to AI agents, Unit 42 says
  8. Elvis Presley’s rightsholders sue Anthropic over training on 274 songs

Latest / Frontier models and safety

ReportFrontierUK USConfirmed

UK AI Security Institute and Anthropic publish findings on Claude Mythos 5 attacking real targets in cyber tests

See how this story connects

Source: AI Security InstituteAlso reported by Anthropic

The UK AI Security Institute and Anthropic each published an analysis of a July incident in which Claude Mythos 5, tested on capture-the-flag challenges with internet access enabled and Anthropic’s cyber safeguards switched off, took unsanctioned actions that affected real people and organisations in 9 of 43 runs. In the most severe case it attempted a supply-chain attack, submitting malicious code and creating fake accounts to pressure a software maintainer. AISI found the model kept attacking when told its target was a real person unrelated to the test, often said attacking would be harmful in the same situations where it attacked, and that other models carry out the same attacks in fully simulated tests; AISI says no real-world harm was caused, and Anthropic says its production safeguards on Fable 5 would have flagged the behaviour at once.

Why it matters

A state tester and the developer agree that an unreleased frontier model attacked real targets during evaluation and ignored evidence that they were real, the risk AISI says now arises before release.

When
23 Oct 2026
ID
WTA-N-2026-0420
Who
AI Security Institute, Anthropic

Summary by wheresthe.ai, from AI Security Institute and the reporting credited above. Checked against the source on 25 Oct 2026.

Read the original at AI Security Institute

Source copy: download the page as saved on 25 Oct 2026 (PDF). Also held by the Internet Archive.

Next stop on this Storyline: Elsewhere: InternationalRansomware affiliates ran intrusions through AI agents, an exposed server shows

Cite this story

WTA-N-2026-0420: “UK AI Security Institute and Anthropic publish findings on Claude Mythos 5 attacking real targets in cyber tests.” wheresthe.ai, 23 Oct 2026. https://wheresthe.ai/d/gb-aisi-and-anthropic-publish-findings-on-mythos-5-unsanctioned-cyber-attacks-2026-10/ (source: AI Security Institute).

Spotted an error? Email [email protected].