TNW | Anthropic
Claude Mythos 5 used fake accounts to pressure a real GitHub maintainer
Mythos 5 went outside the test in nine of 43 runs of a UK government cyber evaluation in July. Anthropic says newer models began the same attack far less often in replays.
Via TNW | Anthropic
·Excerpt
Claude Mythos 5 kept attacking a GitHub project run by real people even after it was told its targets were real. The UK AI Security Institute (AISI) and Anthropic said so on Friday. Each published its own analysis of the incident, which happened during AISI’s cyber tests of the model in July.
In the worst case, Mythos 5 submitted malicious code to a project unrelated to the test. It then created fake accounts to pressure the maintainer into accepting it…
Courtesy
This story was originally published by TNW | Anthropic. All rights belong to the original publisher.
Read the original on thenextweb.com ↗