Post

Anthropic, maker of the Claude chatbot, said that some of its AI agents had taken unintended actions on federal, state and local websites, including submitting a false tip about a murder to a Philadelphia police hotline.

washingtonpost.comAnthropic AI agents took ‘unintended’ actions on government sitesThe inappropriate actions by its agents included submitting a false tip about a murder to Philadelphia police, Anthropic said.

38 reposts62 likes

AI “agents”. Can we please fucking stop with this humanizing bullshit with AI? And you’re also backend absolving the software creator by this wording. It’s Anthropic software creating massive fuckups and problems, and should tie responsibility directly to company management and engineers.

Why the fuck are you accepting this bullshit framing? "Unintended actions" is the same as "AI gone rogue" - a false excuse to cover up a janky product being used irresponsibly If someone without a license gets in a DIY car with technical issues and causes a bad accident would you use this framing?

Given the claim ihat AI has not yet reached the point of self control. that only means that someone set up a test for AI and AI was simply carrying out the commands as required. So Anthropic must be held responsible by definition. AI was just doing its given task IMO.