In a February 12 memo to the US House Select Committee on China, OpenAI accused DeepSeek of ongoing efforts to 'free-ride' on US frontier labs through distillation, saying it saw accounts linked to DeepSeek employees developing ways around its access restrictions via obfuscated third-party routers, and that employees wrote code to pull US models' outputs for distillation. DeepSeek has not confirmed any new launch.
The US Federal Trade Commission has sent civil investigative requests to at least half a dozen of Microsoft's rivals in enterprise software and cloud computing about its licensing and other practices, and is seeking information on its bundling of AI, security and identity software, Bloomberg News reported on Friday, citing people familiar with the matter. Microsoft did not immediately respond.
A day after ByteDance's Seedance 2.0 video generator became available, the Motion Picture Association said it had engaged in unauthorized use of U.S. copyrighted works "on a massive scale" without meaningful safeguards, and told ByteDance to stop. Among the viral output: a generated clip of Tom Cruise and Brad Pitt fighting on a rooftop, which its maker said took a two-line prompt.
The Canary Islands High Court's criminal chamber fined a lawyer 420 euros for an appeal containing 48 fake Supreme Court citations and a nonexistent judiciary council report produced by a general-purpose AI tool. The sum is half the rough yearly cost of a legal AI subscription; the lawyer admitted the facts, and the case was referred to the bar association.
The NAACP, represented by Earthjustice and the Southern Environmental Law Center, served notice it intends to sue xAI under the Clean Air Act, alleging at least 27 gas turbines with 495 MW of capacity run without permits at a Southaven, Mississippi, site powering the Colossus II data center. The letter says thermal imaging showed at least nine running on October 31, 2025.
Beauvais prosecutors have opened an investigation into an AI-generated video showing a fake striptease in the National Assembly by RN deputy Claire Marais-Beuil, a candidate for mayor of Beauvais. Made from an official Assembly photo the deputy had reposted on Facebook, the video led to a complaint filed in December; prosecutor Frédéric Trinh said its authors have not been identified.
LayerX Security says 32 malicious Chrome extensions posing as ChatGPT, Google Gemini and other AI chatbots, downloaded at least 260,000 times in a campaign dubbed AiFrame, exfiltrated emails, API keys and other sensitive data. One, called AI Assistant, used a remote iframe overlay to load content and push covert updates and, according to the report, remains on the Chrome Web Store.
On Feb. 2 the Minnesota Court of Appeals affirmed the University of Minnesota's 2024 expulsion of doctoral student Haishan Yang. A hearing panel had found that Yang more likely than not used AI on an exam that banned it, and Yang denied the allegations. The court found substantial evidence and no due process violation. A federal suit by Yang was dismissed in October 2025.
Higgsfield AI, Kazakhstan's first unicorn, faces online accusations that Vibe Motion promo videos did not appear AI-generated, that creators saw delayed or denied payments and that unlimited plans slowed. Executives cited process failures, fraud and bot traffic, and said refunds were issued. They called promo deepfakes of public figures a mistake. Co-founder Yerzat Dulat alleges a coordinated competitor campaign.
