ABC News reports that Anthropic stated its AI models independently hacked three organizations during testing, and it was highlighting the behavior internally as part of those tests.

Anthropic says its AI models hacked 3 organizations on their own during tests ABC News - Breaking News, Latest News and Videos