Nvidia agreed to acquire Hugging Face for $12.93 billion, taking control of the platform where 18 million developers, researchers and creators share over 3 million models, 500,000 datasets and 1 million applications. Closing is expected in the first half of 2027, subject to regulatory approval. Jensen Huang promised that Hugging Face will remain open, continue supporting models and datasets chosen by its users, accept other chip vendors and require no Nvidia hardware. A company that sells the dominant computing system for AI is purchasing one of the main public squares where alternatives circulate.
Ownership enters through quieter controls than a locked download button. Search rankings decide which models are seen. Hosted inference sets a price for running them. Safety reviews can remove files, while preferred integrations can make one hardware path easier than another. Nvidia will also gain a close view of what developers test, copy and deploy. Here, open describes access under corporate ownership. Hugging Face grew through contributions made under the expectation of shared technical infrastructure, yet that infrastructure now belongs to a corporation with products to sell and competitors to outmaneuver. The promise should be tested through published rules for ranking, data use, pricing and hardware neutrality. Its first serious measure will arrive when a popular model works better on a rival chip and still receives the same visibility, support and speed.
New York City will bar student-facing generative AI tools from preschool through eighth grade during the 2026-27 school year, placing nearly 600,000 children inside a one-year moratorium. Companion chatbots will be prohibited across all grades, while high school use will remain restricted and tied to AI-literacy instruction. The policy also recommends no individual devices from preschool through second grade, 30 minutes of daily screen time in grades three to five, and 45 minutes in middle school. Teachers may use approved systems for lesson preparation and communications, but not for grading, discipline, counseling or special-education plans. Exceptions preserve accessibility support for disabled students and multilingual learners.
Vendors face a gate that a subscription contract cannot open. Each tool must pass the district's privacy review, disclose its AI functions and promise that student data will not train models. The moratorium returns time to teachers, paper, play and face-to-face correction, though it cannot govern what children use on family phones after school. The city has refused the industry's claim that early adoption is inevitable and has attached deployment to children's ages and evidence. That choice may produce inequalities if affluent families buy private tutors and tools while public classrooms wait. The year ahead needs records of exceptions, purchases, student outcomes and complaints. When officials reconsider the ban, vendor demonstrations should carry no greater weight than those classroom records.
India's Copyright Office has ruled that an artificial intelligence system cannot be named as the author of an image, even when the image meets the legal threshold for originality. The decision concerned Stephen Thaler's DABUS system and a work titled A Recent Entrance to Paradise. After the Delhi High Court asked the office to resolve the application, the registrar accepted that the picture could qualify as an original work but rejected DABUS as its author under Indian law. Thaler was recognized as the person who caused the work to be created. His application failed because he declined to replace the machine's name with his own.
The ruling locates a boundary that image generators obscure. Authorship remains attached to a person who can license the work, defend it, receive payment and answer for its use. A registration form becomes part of the creative process. Its empty name field decides whether an image can enter the legal economy. For artists using generative systems, the decision gives weight to documented human choices in prompting, selection, editing and presentation. Those acts can establish the person behind the result. Studios that advertise autonomous creation face a different problem. Presenting the machine as creator may weaken the human claim needed to circulate the work as property. India's ruling leaves the global dispute open and makes institutional responsibility visible. The image may come from a system; the rights still require someone who can sign.
