AI News Roundup – AI regulation moratorium pulled from U.S. Senate bill, Cloudflare to block AI web crawlers by default, Apple considers adopting third-party AI tools, 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.

• • The Associated Press reports on the demise of a proposed moratorium on state-level AI regulation that was originally part of a legislative package before the U.S. Congress. The original form of moratorium, championed by Texas Senator Ted Cruz, would have prohibited states from imposing regulations on AI technologies for a period of 10 years. However, this provision ran into Senate procedural issues, and was eventually modified to only apply as a condition of states receiving grants from a $500 million AI infrastructure development fund, as this roundup reported last week. However, in a 99-1 vote this past week, the Senate removed the provision entirely en route to the passage of the “Big Beautiful Bill Act,” which was signed into law by President Donald Trump. The removal of the provision marks a defeat for the many major AI companies, including OpenAI, the defense startup Anduril, and the venture capital firm Andreessen Horowitz, which had each lobbied heavily in favor of the measure.
• • The MIT Technology Review reports on new measures taken against the web crawlers that comb the Internet for material to use in AI training. Cloudflare, a major Internet infrastructure company, announced that websites using Cloudflare’s network will default to blocking web crawlers from accessing their websites. These web crawlers are programs that access websites to gather information on its content. In the past, they were often used for indexing search engines, but are now also used for obtaining web content for use in AI training. Traditionally, sites would allow search engines to index websites as the websites would then appear in the search results, driving traffic to the site in question. However, with the rise of AI-assisted internet search, such traffic has declined dramatically. Cloudflare claims that the new default (and associated opt-in settings) will allow its clients to have greater control over how AI crawlers access their information, even allowing site owners to whitelist certain crawlers while disallowing others, as well as introducing a “pay-per-crawl” system that allows sites to charge a crawler a certain amount to access the website. Bot access to websites is often regulated by “robots.txt,” a file that provides instructions to bots, but some AI companies have been accused of crawling websites in violation of these instructions. Cloudflare has developed other anti-crawler technology in recent months, including creating a “maze” of fake AI-generated sites to confuse the crawlers, but the company hopes that the new policies will lead to better-faith negotiations between site owners and AI companies.
• • Bloomberg reports that Apple is considering a partnership with OpenAI or Anthropic to power the company’s forthcoming updated Siri voice assistant. Apple’s current AI technologies use in-house models known as Apple Foundation Models, which are planned to be incorporated into new versions of Siri in 2026. Siri, originally released in 2011, has been viewed as falling behind AI-powered chatbots in terms of capabilities, but Apple’s long-planned upgrades have been delayed for several years. Given these challenges, and after leadership turnover in Apple’s Siri team, the company approached both OpenAI and Anthropic for access to their ChatGPT and Claude models (respectively) that could be used for testing on Apple’s servers. Anthropic’s Claude chatbot won out in testing, and Apple is currently in discussions with Anthropic about using Claude to power the “new” Siri. Apple is continuing to develop its in-house AI models, but the possible partnership with other AI companies (or a possible acquisition of such, as Apple has considered acquiring AI startup Perplexity) demonstrates Apple’s desire to charge forward in the AI space.
• • The Wall Street Journal interviews the CEO of the Recording Academy about AI’s effect on the music industry. Harvey Mason, Jr., a former songwriter and producer, said that the threat posed by AI “feels existential,” even in comparison to past technological disruptions in the music industry such as the file-sharing application Napster. Mason said that generative AI will “totally upend the creative process,” and that his greatest concern is how copyright and remuneration will work with generative AI works. Mason reiterated his support for the bipartisan No Fakes Act, which would protect the voice and visual likenesses of individuals (including musical artists) from unauthorized computer-generated copies. As this roundup has previously reported, several music companies (including Universal Music Group, the world’s largest) have sued AI companies like Anthropic on the grounds of copyright infringement. Mason told the WSJ that working with AI companies “makes the most sense,” saying that he does not believe that AI companies are seeking to hurt the music industry, but that “it’s important that we set up guardrails.”
• • Nikkei Asia reports on the usage of AI in video game development. Krafton, South Korea’s largest game developer, has implemented small language models (SLMs) into its most recent simulation game, intended to allow non-player characters (NPCs) within the game to act and respond in a more natural manner. The company partnered with Nvidia to create a model that was small enough to run on users’ devices and still generate responses of an acceptable quality. The game, inZOI, has proved to be popular, selling over 1 million copies in its first week, but the company is worried about AI-in-games competition from Chinese companies, who have also implemented AI-powered features into games. However, some believe that AI features will become commonplace as the technology advances, though most game development work still cannot be done by AI.