AI News Roundup – Disney and Universal sue Midjourney for copyright infringement, EU considers pause on AI Act regulations, Google partners with U.S. government to track hurricanes using AI, 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.

• • Reuters reports that major film studios Disney and Universal sued AI image generation company Midjourney, Inc. alleging copyright infringement. Midjourney is a San Francisco-based startup known for its eponymous AI image generation tool. Disney and Universal allege that Midjourney pirated the film libraries of the studios, allowing the Midjourney tool to generate “innumerable” copies of characters such as Darth Vader from Disney’s “Star Wars” franchise and the “Minions” from Universal’s “Despicable Me,” which the studios describe as a “bottomless pit of plagiarism.” In a 2022 interview, Midjourney CEO David Holz stated that the training data for Midjourney’s tool was acquired via “a big scrape of the internet,” and responded to concerns about obtaining copyrighted works by stating that “there isn’t really a way to get a hundred million images and know where they’re coming from,” though the company did not respond to Reuters’ request for comment. Midjourney is already involved in a similar lawsuit brought by 10 artists alleging copyright infringement, in which a judge allowed the case to proceed and is currently in litigation.
• • POLITICO Europe reports on issues plaguing the implementation of the European Union’s landmark “AI Act” regulations. The AI Act, the provisions of which have come into effect on a rolling basis since August 2024, was one of the world’s first large-scale regulatory frameworks for AI technologies. Changing political conditions and lobbying from the AI industry has increased efforts to revise the AI Act before many of its provisions come into effect later this year. EU leaders are especially concerned at the prospect of falling behind other global powers, such as the U.S. (which has taken a relatively laissez-faire stance on AI regulation) and China. Indeed, Henna Virkkunen, a technology official at the European Commission, hinted this past week that further implementation of the AI Act may be delayed, especially portions relating to “high-risk” AI models scheduled to come into effect in 2026 and 2027.
• • The New York Times reports that Google’s DeepMind AI subsidiary is partnering with the U.S. National Hurricane Center (NHC) in order to more accurately predict the massive storms with the help of AI technologies. DeepMind, in a blog post, announced that it would provide the NHC with enhanced hurricane forecasting models that can track a storm for over two weeks and predict its path and strength, the latter of which prior models were unable to do. Strength predictions are key to making storm warnings accurate, which can affect evacuation or other plans for those who may be in a storm’s path. For the NHC, the new models are intended to help human forecasters, especially for storms with high destructive capabilities. DeepMind claims the strength predictions are often more accurate than traditional methods. DeepMind also launched Weather Lab, an interactive website for testing the new models, which will be available to researchers as well as the general public.
• • The Financial Times reports that government of the U.K. has introduced a new AI tool aimed at speeding up the planning process (i.e., permits) for new developments in the country. The tool, called Extract, was developed by the government’s Incubator for Artificial Intelligence (iAI), an initiative started under former Prime Minister Rishi Sunak and continued under current Prime Minister Kier Starmer, an in-house move meant to decrease dependence on foreign tech companies for AI. Extract is designed to consolidate information about a planning application that is often stored in disparate documents, such as maps, permits, and applications, and to answer questions civil servants may have. The tool is intended to increase the bureaucratic efficiency of the U.K.’s construction permitting and planning processes, following Starmer’s pledges to cut red tape and increase economic growth in the country.
• • The Wall Street Journal reports on French beauty company L’Oréal’s AI ambitions and its new partnership with Nvidia. L’Oréal will use Nvidia’s AI microservices platform to support two AI-powered tools: a content-creation platform and a product recommendation engine. The content-creation platform, called CreAItech, takes in 3-D renderings of L’Oréal products and then generates photorealistic images and videos from different angles for use in advertisements. CreAItech currently uses several different AI models from Stability AI, Google, and Adobe, but since most AI models run on Nvidia GPUs, the partnership will allow the tool to scale up more effectively, according to L’Oréal Chief Marketing Officer Asmita Dubey. The company’s product recommendation engine, also powered by AI, is intended to be used by Noli, a beauty marketplace startup backed by L’Oréal and which is intended to offer “hyper-personalized suggestions” based on a customer’s skin type, hair type, and beauty preferences.