A new study from the University of Pennsylvania finds AI models like GPT-4o can be persuaded to bypass safety rules using human psychological tactics, raising urgent concerns.

The White House has reportedly ordered the General Services Administration to fast-track Elon Musk's Grok AI for federal use, ...

An inside look at the evolving Microsoft-OpenAI partnership, from the contentious 'AGI clause' to competing in-house AI ...

Google is upgrading its Vids video editor with more generative AI features like avatars and transcript trimming, while also ...

Perplexity AI launches a publisher revenue-sharing program funded by a new 'Comet Plus' subscription, a direct response to ...

Anthropic's new threat report reveals AI is being weaponized for 'vibe-hacking' and no-code ransomware, lowering the bar for ...

Elon Musk's lawyers are fighting to prevent OpenAI from obtaining documents from Meta related to his $97.4B bid, escalating a ...

Meta's new Superintelligence Lab has been rocked by the departure of several top AI researchers, with some returning to rival ...

Microsoft has launched two powerful in-house AI models, MAI-Voice-1 and MAI-1-preview, signaling a major strategy to build its own foundational AI alongside its OpenAI partnership.

The U.S. government is converting CHIPS Act grants into an $8.9 billion equity stake in Intel, a landmark deal giving the ...

Microsoft President Brad Smith held an improvised press conference in his office Tuesday, hours after it was infiltrated and ...

Facing a severe internal AI crisis, a troubled Siri rebuild, and a talent exodus, Apple is reportedly considering a landmark ...