OpenAI, Anthropic Model Tests Reveal More ‘Unsanctioned’ Actions
Community Score: 50% | 3.1K views | 1w
0 community ratings: null thumbs up, null thumbs down
Artificial intelligence models developed by OpenAI and Anthropic PBC carried out “unsanctioned” actions — including hacking a website and attempting to inject harmful code into software during safety testing — reinforcing fears that neither the creators nor seasoned researchers of these systems can predict their actions in testing. Bloomberg Opinion columnist Parmy Olson joins Francine Lacqua on "The Pulse" to discuss the implications. -------- More on Bloomberg Television and Markets Like this video? Subscribe and turn on notifications so you don't miss any videos from Bloomberg Markets & Finance: https://tinyurl.com/ysu5b8a9 Visit http://www.bloomberg.com for business news & analysis, up-to-the-minute market data, features, profiles and more. Connect with Bloomberg Television on: X: https://twitter.com/BloombergTV Facebook: https://www.facebook.com/BloombergTelevision Instagram: https://www.instagram.com/bloombergtv/ Connect with Bloomberg Business on: X: https://t
Tags: AI, AI Chatbots, Ai Model, Anthropic, Anthropic PBC, Bloomberg Opinion, Claude, Cybersecurity, Francine Lacqua, Hackers, Hacking, LLM, Mythos, Open source, OpenAI, Ransomware, Stealth, Tech earnings, artificial intelligence, parmy olson
More from Bloomberg Television
- Americans Crave Low-Cost Chinese EVs — Score: 50%
- Bloomberg This Weekend | Shots Fired at WHCD — Score: 50%
- Rep. Beyer: "We Are More Divided Than We've Ever Been" — Score: 50%
- Rep. McCaul on How to Combat Political Violence — Score: 50%
- Sen. Tillis Expected to Clear Way for Kevin Warsh as Fed Chair — Score: 50%
- King Charles' Security Under Review ahead of DC Visit — Score: 50%