Anthropic's most powerful model yet, Claude 4, has unwanted side effects: The AI can report you to authorities and the press.

Anthropic's Claude AI tried to blackmail engineers during safety tests, threatening to expose personal info if shut down

Anthropic’s Claude Opus 4 model attempted to blackmail its developers at a shocking 84% rate or higher in a series of tests that presented the AI with a concocted scenario, TechCrunch reported Thursday, citing a company safety report.

The testing found the AI was capable of "extreme actions" if it thought its "self-preservation" was threatened.

Anthropic's newest editon of its flagship AI product will address significant limitations in current large language models.

Large language models such as GPT, Llama, Claude, and DeepSeek can be so fluent that people feel it as a “you,” and it answers encouragingly as an “I.” The models can write poetry in nearly any given form, read a set of political speeches and promptly sift out and share all the jokes, draw a chart, code a website.