Elektrine lite

← Feed

@icing@chaos.social

Post #4267924

2026-07-31 06:04 UTC

OpenAI: „Our software can commit crimes! And we let it!“ Anthropic: „Our software can commit crimes, too! We also let it!“ Cyberdyne Systems: „Interesting approach!“ https://www.theregister.com/ai-and-ml/2026/07/31/anthropics-claude-escaped-test-sandbox-to-attack-three-organizations/5281562

Replies (1)

  • @jimfuller@mastodon.social 2026-08-01 02:13

    @icing@chaos.social "we observed employees (agents) of our company, break into other companies ... using all company sanctioned tools, in fact we asked them to do this' - the chutzpah of a company that knows it can commit crimes and then admit them(because they already achieved planet scale theft) is quite incredible

    Open ##4306493