To help you stay on top of the latest news, our AI practice group has compiled a roundup of the developments we are following.
- The Trump administration has moved quickly on AI in its first week in office, revoking former President Biden’s executive order that sought to establish guidelines for the rapidly evolving technology, according to the Associated Press. Biden’s order, signed in October 2023, required AI developers to share AI safety test results with the government and directed the National Institute of Standards and Technology (NIST) to create AI standards, as well as directed the Department of Commerce to establish guidelines for watermarking AI-generated content. Hours after taking office, President Trump revoked this order, fulfilling a campaign promise. Later in the week, Trump issued his own executive order relating to AI, directing several White House officials to develop an “Artificial Intelligence Action Plan” within the next 180 days, and stating that the U.S. “can solidify [its] position as the global leader in AI.” Further action on AI from the Trump administration is expected over the next few months, especially as the leaders of large tech companies, including Elon Musk, Meta CEO Mark Zuckerberg, Amazon CEO Jeff Bezos, and Apple CEO Tim Cook have courted favor with the new president and were guests at Trump’s inauguration.
- Chinese AI company DeepSeek recently released DeepSeek-V3 and its counterpart reasoning model DeepSeek-R1, claiming performance exceeding that of current AI leaders, according to The New York Times. According to a paper accompanying the new models, DeepSeek-V3 matched or outperformed the latest models from American AI leaders such as OpenAI, Anthropic, and Meta in several standard industry benchmarks. DeepSeek engineers claimed that the new models were trained on only around $6 million worth of computation power, far less than the resources used to develop competitors’ models. One expert told the NYT that due to the U.S.’ restrictions on the most advanced AI chips being exported to China, it forced developers in that country to train models more efficiently on the hardware they already had. DeepSeek, a start-up division of the quantitative trading firm High-Flyer based in Hangzhou, China, does not make consumer products, and focuses its AI work on open research; indeed, DeepSeek’s models and weights were released under the permissive MIT software license, allowing others to use the models without royalties. These new developments demonstrate how quickly competitors have been able to catch up to AI leaders and heralds greater competition in the industry, especially from outside the U.S.
- OpenAI, Oracle, Softbank, and the Trump administration have announced “The Stargate Project,” a joint venture between the three companies intended to build AI infrastructure, according to The Washington Post. OpenAI CEO Sam Altman, Oracle Chairman Larry Ellison and SoftBank CEO Masayoshi Son joined President Trump in the White House to announce the project, which will aim to spend as much as $500 billion over the next four years to construct AI datacenters around the country. These facilities, which often contain specialized hardware, are key to the development and running of powerful AI models. According to OpenAI, technology processor company Arm, along with Microsoft, and Nvidia will provide hardware and software for the project, which Altman described as “the most important project of this era.” Such facilities also demand enormous amounts of electricity, beyond what the U.S. power grid is currently able to produce. Several tech CEOs, including Altman and venture capitalist Marc Andreessen, have invested in nuclear power companies and efforts to reactivate dormant nuclear plants in efforts to use clean energy for AI, while Trump’s interior secretary nominee, former North Dakota Governor Doug Burgum, called for investments in the extraction of coal and natural gas to power AI development.
- During this year’s winter X Games held in Aspen, Colorado, an AI system will be used for the first time in judging a snowboarding competition, according to The New York Times. During the SuperPipe snowboard event, cameras will capture the spins, flips and grabs performed by the athletes, while an AI system will use that information to produce a score for the routine. It is a test-run for the technology; human judges will still provide the official scores, but the AI-provided score will also be available for audiences. Sports are no stranger to technology, but AI has accelerated changes in the judging of sports. Line judges at the Wimbledon tennis championships will be replaced by an AI-powered system starting this year, while the AI-based Judging Support System has been used since 2023 to aid judges at the gymnastics World Championships. Proponents of AI judging say that the technology has the potential to reduce bias, especially in subjectively scored sports such as gymnastics and figure skating. The X Games AI implementation was welcomed by judges as helping to provide, in addition to scores, further insights into tricks performed by athletes. Jeremy Bloom, chief executive of X Games, told the NYT that “sometimes you get the tricks wrong because they’re spinning so fast…[but] this model, because it can watch the video in slow, slow motion, is really accurate in its ability to say, ‘Well, that was a cab 1400.’”
- A leading contender for Best Picture at this year’s Academy Awards has received backlash after editors revealed that AI had been used in the production of the film. The Brutalist, starring Adrien Brody as a Hungarian Jewish Holocaust survivor who comes to America and becomes a leading architect, was revealed to have used an AI tool to improve the acted Hungarian accents of Brody and actress Felicity Jones in the film. Brody and Jones recorded their dialogue into the tool, known as Respeecher, which modified the voices based on the accent of Hungarian-born editor Dávid Jancsó. Jancsó also said that generative AI was used as a basis for a “series of architectural drawings and finished buildings” that appear in one sequence in the film. These revelations caused immediate backlash on social media, with one user calling the use of AI a “disgrace” to manipulate actor’s voices, which was “a fundamental aspect of…acting,” as well as for using the generative AI to “avoid paying visual artists for their work.” In response, the film’s director, Brady Corbet, said in a statement to The Hollywood Reporter that “Adrien and Felicity’s performances are completely their own… [i]nnovative Respeecher technology was used in Hungarian language dialogue editing only, specifically to refine certain vowels and letters for accuracy,” and that the “aim was to preserve the authenticity of Adrien and Felicity’s performances in another language, not to replace or alter them and done with the utmost respect for the craft.” With regard to the generative AI, Corbet told THR that while generative AI was used as inspiration for images used in the sequence in question, “[The Brutalist production designer] Judy Becker and her team did not use AI to create or render any of the buildings. All images were hand-drawn by artists.” The furor over even small uses of AI technologies in filmmaking demonstrate that their use remains a sensitive topic, especially as copyright infringement lawsuits continue against developers of AI systems trained on art and videos available online without permission from their creators.
[View source.]
