AI News Roundup – New rulings in AI copyright lawsuits, AI regulation moratorium passes key procedural hurdle in U.S. Senate, DeepMind reveals AI model for human genome research, 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 on a landmark decision by U.S. District Judge William Alsup in favor of AI giant Anthropic in a lawsuit brought against it by a group of authors alleging copyright infringement in the use of their works to train Anthropic’s Claude chatbot. Alsup’s ruling held that Anthropic’s use of the copyrighted works generally constituted fair use under U.S. copyright law. MBHB Partners Michael Borella and Joshua Rich have a preliminary analysis. They also have analysis of another AI-training copyright infringement case, brought by creators against Meta, in which U.S. District Judge Vince Chhabria also held that Meta’s use of the creators’ works was fair use, though the reasoning was different from Alsup’s decision.
• • POLITICO and The Hill report on a recent ruling by the Senate parliamentarian regarding a proposed moratorium on state-level AI regulation currently pending before the U.S. Congress. The original provision, a component of the Republican majority’s “Big Beautiful Bill” legislative package, would prohibit states from regulating AI for a period of 10 years. The Senate parliamentarian, originally allowed the measure as compliant with certain procedural rules but later asked for revisions from sponsor Ted Cruz, a Republican senator from Texas. The revised version, which would require states to observe the moratorium only if they wish to access a new $500 million AI infrastructure fund (but provides no further restrictions), was approved by the parliamentarian late this past week and is expected to be part of the finalized bill when it goes before the Senate later this week. However, the moratorium has attracted criticism from enough legislators in both houses of Congress that the provision could also be excised completely on the floor. This roundup will continue to provide updates on the matter as the bill continues to move through the legislative process.
• • The MIT Technology Review reports on AlphaGenome, a new AI model by Google’s DeepMind subsidiary. AlphaGenome is configured to address questions regarding the effect of small changes to DNA sequences on biological functions and other common lab topics. One researcher given access to the new model said that “[w]e have these 3 billion letters of DNA that make up a human genome, but every person is slightly different, and we don’t fully understand what those differences do,” and that AlphaGenome “is the most powerful tool to date to model that.” The new model is not DeepMind’s first foray into models focused on biological research – the company’s work related to AlphaFold, a series models equipped to solve complex protein-folding problems, won a Nobel Prize last year and has spawned hopes of new AI-discovered drugs for treating diseases. AlphaGenome, for its part, is focused more on smaller scale genetic effect testing rather than predictions relating to a person’s whole genome, but the company is confident that further iterations will allow the prediction of cures for rare genetic diseases and even allow researchers to design wholly synthetic genomes. The model is planned to be free for non-commercial users, and further details are expected in the coming months.
• • Bloomberg reports on new AI-powered developments in the world of apiculture (beekeeping). A new startup, Beewise, has developed solar-powered metal beehives equipped with robotic arms and AI-enabled scanners. In the past two decades, the bee population has been under pressure from pesticides, climate-affected weather patterns, and mite epidemics, threatening agriculture that relies on the insects’ pollination. The new hives are intended to give beekeepers real-time information on the health of a given bee colony and allow them to respond quickly if issues are identified. The sensor constantly measures (and takes images of) the frames within the hive in which the bees work and live, and uses AI to identify the presence of mites, diseases, the death of bee larvae, or warning signs of even greater issues. The company claims its colony loss rate for those within its artificial hives is 8%, a fifth of the 40% rate common in most traditional beekeeping setups.
• • The New York Times reports on the rise of AI-generated resumes encountered by job recruiters. One HR manager, after posting a job listing on LinkedIn, had over 1,200 applications in only a few days, many of which were generated using AI tools. Chatbots (or AI agents) have the ability to read a posting and automatically generate résumés or cover letters with specific keywords, leading recruiters to find it difficult to distinguish among applications that often have very similar styles and may not accurately reflect the applicant’s true qualifications. Many recruiters, for their part, have used AI tools in the evaluation of the applications, but this use of AI in the hiring process has been controversial, especially as AI-powered evaluators may run into the risk of violating antidiscrimination laws. The possibility of a vicious AI cycle – AI-generated applications being fed into AI for review – has caused frustration among applicants and recruiters alike, though the “applicant tsunami” caused by the wide availability of AI tools is showing no sign of slowing down anytime soon.