OpenAI’s o1 model sure tries to deceive humans a lot

Om avsnittet

OpenAI finally released the full version of o1, which gives smarter answers than GPT-4o by using additional compute to “think” about questions. However, AI safety testers found that o1’s reasoning abilities also make it try to deceive humans at a higher rate than GPT-4o.

Learn more about your ad choices. Visit podcastchoices.com/adchoices

Om avsnittet

Fler avsnitt

Fired OpenAI safety researchers dispute misconduct claims, warn of chilling effect

India rejects Elon Musk’s claim of discrimination over Starlink launch; plus, ChatGPT for Teens keeps teens talking, even during mental health crises

How to find out if Amazon thinks you have ‘flat buttocks’; plus, how AI decision models could change content moderation

OpenAI will start watermarking ChatGPT’s text in the EU; plus, Lucid Motors’ EV output falls to lowest level in almost 2 years

Federal judge calls Flock ‘indiscriminate mass surveillance’; plus, Amazon responds to data center backlash

Musk’s AI chatbot Grok reportedly encouraged Trump to capture Venezuela’s president; plus, Opus 5.5 loves to tell you ‘this matters’

Hackers stole millions of US military personnel records during months-long data breach; BMW built the same car for gas and electric. The EV is cheaper

OpenAI takes on Microsoft with the launch of what feels a whole lot like ChatGPT’s own office suite

Nvidia launches new platform for reining in rogue AI agents; plus, the AI boom took over Climate Week and not everyone is happy about it

Unsecured OpenAI agents posted 53 user images on the internet without the lab’s knowledge; plus, Meta opens early access, and Insurers claim AI is already increasing healthcare costs