AI News Roundup – New executive order targets state AI laws, Trump administration permits advanced Nvidia AI chip sales to China, Disney inks licensing deal with OpenAI, 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.

• • A new executive order signed by President Trump this past week directs the federal government to target state-level AI laws, according to POLITICO. The measure, which broadly mirrors a leaked version of the order that this AI Roundup covered last month, is intended to “ensure that there is a minimally burdensome national standard” for AI regulation. To that end, the order creates an “AI Litigation Task Force” within the U.S. Department of Justice to challenge state AI laws that “unconstitutionally regulate interstate commerce, are preempted by existing Federal regulations, or are otherwise unlawful in the Attorney General’s judgment.” Further, it directs the Department of Commerce to evaluate and identify “onerous” state AI laws, and to withhold federal broadband funds from states with such laws. However, the released order differs from the leaked one in a key respect: the order does not call for a federal law that preempts state AI regulations that protect children’s safety. As reported by this AI Roundup last week, a push in Congress for such a preemption law had stalled in recent weeks amid pushback within the Republican Party, likely prompting the current executive order. Indeed, several Republican governors, including Florida’s Ron DeSantis and Utah’s Spencer Cox, have pushed ahead with their own AI plans and have objected to any federal efforts at preemption. Governors of the opposition Democratic Party, such as California’s Gavin Newsom, criticized the order as a “con,” while other Democrats promised legal challenges to the order.
  • The New York Times reports that the Trump administration has greenlit sales of Nvidia’s advanced H200 AI chips to China. Exports of H200 chips, Nvidia’s second-most-powerful offering behind its industry-leading Blackwell line, to China had previously been restricted by the U.S. Commerce Department. The move is a major victory for Nvidia CEO Jensen Huang and White House AI czar David Sacks, who have spent months lobbying against export controls, arguing that allowing sales could make China dependent on U.S. technology, especially as China has attempted to promote the use of domestic AI chips like those from Huawei and Cambricon. However, the weakening of export controls to China has proven unpopular in the U.S. Congress, as a bipartisan group of senators has proposed legislation that would further restrict AI chip sales to China (as this AI Roundup reported last week). An Nvidia spokesman told The Times that the new measure struck “a thoughtful balance that is great for America” by permitting sales of the chips to Commerce Department-vetted customers in China.
  • The Walt Disney Company has struck a deal with OpenAI to allow the use of the media giant’s content in OpenAI’s Sora generative AI video platform, according to The Wall Street Journal. Sora, launched earlier this year, has been controversial for its ability to generate content containing famous characters who are under copyright. OpenAI initially had an opt-out policy for inclusion in Sora, but backtracked after criticism. Since then, it has sought licensing deals with media companies, and the latest agreement with Disney will allow Sora users to generate videos of over 200 Disney, Marvel, Star Wars, and Pixar characters, though under restrictions on what the characters may do in the videos. In particular, Sora will limit Disney characters in situations related to drugs, alcohol, and sex, addressing concerns about OpenAI’s upcoming “adult mode.” OpenAI will compensate Disney for the use of its characters, though the exact financial terms of the agreement were not made public. Disney will also invest $1 billion into OpenAI and access to purchase further OpenAI stock at its current valuation. Disney CEO Bob Iger said on CNBC that “in this deal, OpenAI is both respecting and valuing our creativity.”
  • Anthropic is donating its Model Context Protocol (MCP), an open standard that facilitates communication between AI agents and outside tools, to the Linux Foundation, according to The Verge. The MCP was released in December 2024, and was intended to create a standard architecture for AI agents to access external tools and data sources, akin to a new type of application programming interface (API) that are commonplace in computing. It has been widely adopted across the AI industry, including by Anthropic rivals OpenAI and Google. Now, the company is donating MCP to the Linux Foundation, a nonprofit that supports the development of the eponymous open-source operating system that is widely used by software developers. Anthropic is also joining with OpenAI, Google, Microsoft, Amazon Web Services, and other corporations to establish the Agentic AI Foundation (AAIF) to “advance open-source agentic AI.” Jim Zemlin, CEO of the Linux Foundation, told the Verge that he was shocked by the rapid adoption of the MCP, and that “when everybody is working collectively on a standard, to help improve the technology so that it’s more secure, so that it can do effective trusted payments, so that it has a better way of communicating in a secure fashion, that’s when the market really gets made.”
  • BBC News reports that Amazon Prime Video has pulled an AI-generated recap for a popular television series after it made several high-profile mistakes. Prime Video recently introduced a generative AI feature to help viewers catch up on shows, but one such recap for the show Fallout, based on the popular video game franchise, made several mistakes, including describing a character as a “1950s flashback” despite the show taking place in 2077, as well as incorrectly summarizing the dynamic between two key characters. The Fallout situation is the second AI-related controversy to involve Prime Video in recent weeks. Earlier this year, Amazon announced that it would use AI to dub certain movies and television series into English and Spanish. Late last month, the AI dubs appeared on several Japanese anime series, and were roundly criticized for “robotic” and “emotionless” voices. Several voice actors also raised concerns about being replaced by generative AI, and some of the AI dubs were eventually removed from Prime Video. Generative AI-induced errors have been widely reported ever since the beginning of the AI boom, especially those with comic effect, though recent high-profile mistakes may cause some companies to reconsider their rapid rollout of AI into their products.