When DeepSeek, Alibaba, and other Chinese firms released their AI models, Western researchers quickly noticed they sidestepped questions critical of the Chinese Communist Party. U.S. officials later confirmed that these tools are engineered to reflect Beijing’s talking points, raising concerns about censorship and bias. American AI leaders like OpenAI have pointed to this as justification…
Demis Hassabis, CEO of Google’s AI research organization DeepMind, appeared to suggest Tuesday evening that Veo 3, Google’s latest video-generating model, could potentially be used for video games. In response to a post on X beseeching Google to “Let me play a video game of my veo 3 videos already,” and asking, “playable world models…
Startups working on AI-powered e-commerce tools often rely on external data or user signals to build and improve their models. However, a company called Remark is taking a different approach by letting thousands of human experts chat with users while they are purchasing items, and then using that knowledge to train models. Remark also makes…
Meta has hired a highly influential OpenAI researcher, Trapit Bansal, to work on its AI reasoning models under the company’s new AI superintelligence unit, a person familiar with the matter tells TechCrunch. OpenAI spokesperson Kayla Wood confirmed to TechCrunch that Bansal had departed OpenAI. Bansal’s LinkedIn page says that he left OpenAI in June. Bansal…
A federal judge sided with Meta on Wednesday in a lawsuit brought against the company by 13 book authors, including Sarah Silverman, that alleged the company had illegally trained its AI models on their copyrighted works. Federal Judge Vince Chhabria issued a summary judgment — meaning the judge was able to decide on the case…
Measuring AI progress has usually meant testing scientific knowledge or logical reasoning – but while the major benchmarks still focus on left-brain logic skills, there’s been a quiet push within AI companies to make models more emotionally intelligent. As foundation models compete on soft measures like user preference and “feeling the AGI,” having a good…
Several weeks after Anthropic released research claiming that its Claude Opus 4 AI model resorted to blackmailing engineers who tried to turn the model off in controlled test scenarios, the company is out with new research suggesting the problem is more widespread among leading AI models. On Friday, Anthropic published new safety research testing 16…
OpenAI researchers say they’ve discovered hidden features inside AI models that correspond to misaligned “personas,” according to new research published by the company on Wednesday. By looking at an AI model’s internal representations — the numbers that dictate how an AI model responds, which often seem completely incoherent to humans — OpenAI researchers were able…
French AI lab Mistral is getting into the reasoning AI model game. On Tuesday morning, Mistral announced Magistral, its first family of reasoning models. Like other reasoning models — e.g. OpenAI’s o3 and Google’s Gemini 2.5 Pro — Magistral works through problems step-by-step for improved consistency and reliability across topics such as math and physics….
Windsurf, the popular vibe coding startup that’s reportedly being acquired by OpenAI, said Anthropic significantly reduced its first-party access to its Claude 3.7 Sonnet and Claude 3.5 Sonnet AI models. Windsurf CEO Varun Mohan said in a post on X that Anthropic gave Windsurf little notice for this change, and the startup now has to…
