AI News Roundup – Chinese companies ready orders for Nvidia AI chips, advertisements coming soon to ChatGPT, Hollywood actors launch anti-AI campaign, and more

To help you stay on top of the latest news, our AI practice group has compiled a roundup of the developments we are following.

• • China’s government has directed major Chinese tech companies to prepare orders for Nvidia’s H200 AI chip, according to Bloomberg. Chinese customs officials had previously blocked shipments of the H200 into the country, according to the Financial Times, despite recent U.S. approval of the sales in December 2025, causing uncertainty for Nvidia’s parts suppliers. However, Chinese regulators have granted “in-principle approval” for Alibaba, Tencent, and ByteDance, three of China’s largest tech companies, to move forward in the process to gain permission to import H200 chips. The Chinese government is likely to condition such an approval on the companies purchasing a certain amount of domestically produced chips as well, as the country seeks to bolster its own semiconductor industry. The H200 is of an older generation of Nvidia’s AI accelerator chips, but has become a flashpoint in trade negotiations between the U.S. and China. The move demonstrates that the Chinese government is willing to grant access to U.S. chips as part of a broader AI and electronics economic strategy. 
  • WIRED reports that OpenAI will soon begin testing advertisements in ChatGPT. The company will begin to roll out ChatGPT advertisements in the U.S., saying that the ads will appear in separate, labeled boxes from the AI system’s usual outputs, and that the ads will not affect those outputs. One example given was, if using ChatGPT to plan a trip, a user may also be given an ad for hotels in the area, using conversation data to match relevant ads. However, the company did not specify exactly what sort of data is collected from users for advertising purposes. The ads first began to appear in ChatGPT late last week for users of the chatbot’s free and Go tiers, but OpenAI said that subscribers to its Plus, Pro, and Enterprise plans will not be served ads. The chatbot, which reportedly has over 800 million weekly users, is one of the largest platforms on the internet, and thus provides an enormous marketing opportunity as OpenAI seeks to generate revenue beyond what it has raised from investors amid competition from AI rivals like Google Gemini. 
  • Hundreds of Hollywood actors have signed on to a new campaign criticizing AI companies’ use of copyrighted data for model training, according to Variety. Over 700 actors and writers, including Scarlett Johansson, Cate Blanchett, and Joseph Gordon-Levitt have backed the “Stealing Isn’t Innovation” campaign, which calls out AI companies for “using American creators’ work to build AI platforms without regard for copyright law.” The use of copyrighted materials to train AI models has been highly controversial, and subject to much litigation in the past few years, though there has yet to be a definitive ruling on whether such use qualifies as “fair use” under U.S. copyright law. While arguing this in court, many AI companies have struck licensing deals with media rights holders for access to their content for training purposes, which the new campaign encourages, saying that “[w]e can have advanced, rapidly developing AI and ensure creators’ rights are respected.” 
  • The MIT Technology Review examines new research projects in the U.K. developing AI systems that can run scientific experiments. The U.K. government’s Advanced Research and Invention Agency (ARIA), recently ran a competition that solicited research proposals for projects relating to “AI scientists,” which are defined by the agency as “a system that can run an entire scientific workflow, coming up with hypotheses, designing and running experiments to test those hypotheses, and then analyzing the results.” Some such systems can even feed results back into the experimental setup to collect more data. In such a system, humans provide a research question which the AI scientist then looks into. ARIA chose 12 projects out of 245 proposals, and each will receive around £500,000 to develop and run their AI scientists for nine months, after which the teams will demonstrate their capabilities and findings. ARIA’s Chief Technology Officer, Ant Rowstron, tried to temper expectations of the systems, saying that he is “not expecting them to win a Nobel Prize,” but that he hopes these projects will lay the groundwork for future scientific advancement and automation. 
  • U.S. retailers are rapidly adopting AI technologies in their sales experiences, according to The New York Times. At the National Retail Federation conference this past week, AI companies openly vied for the attention of retailers, with a large focus on “agentic AI.” Shopping AI agents, which can assist customers and act on their behalf, have rapidly become a mainstay of the online shopping experience. Walmart’s AI assistant, Sparky, can even help customers plan events and purchase items for them. Some retailers are using AI to assist the production of goods, including the luxury goods conglomerate LVMH, which said its employees use AI to visualize final products and colors. AI in the luxury sector has been controversial, with many employees worrying about possible replacement by AI, but an LVMH official said that AI “is a tool that helps everyone to be better” and that the company is trying to “embrace this topic of AI creativity.” Other retailers, including Tractor Supply and Home Depot, however, have concerns over AI errors that could cause customer harm with recommendations or product demonstrations. Despite this, optimism in the industry for adopting the technology is still very high, and further AI integration into the retail experience is expected in the coming year.