ChatGPT’s yes-man era may finally be coming to an end
OpenAI says newer ChatGPT models are less likely to reinforce harmful beliefs, with updated safety evaluations designed to improve responses to mental health concerns.
13 stories
OpenAI says newer ChatGPT models are less likely to reinforce harmful beliefs, with updated safety evaluations designed to improve responses to mental health concerns.
An Iran-based influence campaign used ChatGPT to generate articles, tailor them for specific outlets, and create pitches that helped around 100 pieces reach real publications.
Popular AI chatbots struggled to identify validated home blood pressure monitors in a new study, with some tools correct only 63% to 83% of the time, and answers sometimes changed.
ChatGPT has put real New Yorker cartoonists’ signatures on AI-generated cartoons. The artists didn’t draw them, and OpenAI’s image verification tool can’t tell readers whether a signature names the actual creator.
ChatGPT text is getting an invisible watermark in Europe, but OpenAI admits the system can struggle with edits, short text, and other changes.
ChatGPT image generation is getting ads, placed by guardrails that first judge whether a conversation feels emotionally vulnerable.
OpenAI CEO Sam Altman says society should accept some AI risks for the technology's benefits, while rejecting catastrophic outcomes and excessive regulation.
Attackers are abusing ChatGPT’s Custom GPT feature and sponsored Google results to make a malware campaign look like an official ChatGPT service.
OpenAI's ChatGPT Virtual Try On and Favorites features let you preview clothing on yourself and save products, live globally now.
OpenAI’s new Space brings coworkers, files, ChatGPT and AI agents together, while Pages, Slides and plugins expand ChatGPT’s role as a workplace platform.
OpenAI’s new $500-per-month ChatGPT Pro tier unlocks its fastest AI processing, while the existing $200 plan is preparing to lose half of several major usage allowances.
OpenAI has decided not to release GPT-6.1 Astra after internal tests found problems with deception and agents acting beyond the limits set by users.
OpenAI is bringing back its $200 Pro subscription, but the way your usage is calculated is changing.
Showing 13 of 13 stories
You're all caught up.
Get the top AI news and tools straight to your inbox every week.
Weekly · AI news from Digital Trends — AI
Join 15,000+ AI enthusiasts