U.S. AI fights shift to courts and liability
The U.S. regulatory fight over AI is fracturing into lawsuits and liability battles rather than only policy debate, with Elon Musk’s xAI suing Colorado to block enforcement of that state's 2024 AI law and OpenAI backing an Illinois bill aimed at limiting certain lawsuits over “critical harm.” Those moves signal a patchwork approach where companies test both litigation and legislative defenses against differing state rules. That matters because firms deploying AI will face a mix of state-level requirements and shifting liability regimes rather than a single federal standard. (thehill.com (qz.com))
The fight over artificial intelligence in the United States just moved out of hearing rooms and into courtrooms. On April 9, Elon Musk’s company xAI sued Colorado in federal court to stop a state artificial intelligence law before it takes effect on June 30. (usnews.com) Colorado’s law is not a general “be careful with artificial intelligence” rule. Senate Bill 24-205 targets “high-risk” systems used in decisions about jobs, housing, education, health care, lending, insurance, legal services, and government services. (leg.colorado.gov) (coag.gov) The basic idea is simple: if a software model helps decide whether you get an apartment, a loan, or a job interview, Colorado wants the company behind it to use “reasonable care” to prevent algorithmic discrimination. The law also requires notices, impact assessments, public summaries, and disclosures to the attorney general in some cases. (leg.colorado.gov) (coag.gov) xAI says those rules cross a constitutional line. In its suit, the company argues Colorado is restricting how developers design models and compelling speech through required disclosures, which xAI says violates the First Amendment. (usnews.com) That case is one front in the new battle. Another front is Illinois, where Senate Bill 3444 would limit when developers of “frontier” artificial intelligence models can be sued for “critical harms” if they did not act intentionally or recklessly and if they publish safety and transparency reports. (ilga.gov) (qz.com) Illinois wrote “critical harms” to mean disasters on a very large scale, not ordinary product complaints. The bill summary says it covers events like mass casualties, at least $1 billion in property damage, or the use of a model in creating chemical, biological, radiological, or nuclear weapons. (qz.com) (ilga.gov) OpenAI is backing that Illinois bill. Quartz, citing Wired’s reporting, said OpenAI argued for a system focused on the most advanced models and warned against a patchwork of inconsistent state requirements. (qz.com) Put those two moves together and you get the new map of the fight. In Colorado, a company is trying to knock out a state rule in court; in Illinois, a company is trying to shape the liability rules before lawsuits arrive. (usnews.com) (qz.com) This is happening because Congress still has not passed a single national law that clearly answers who is responsible when artificial intelligence causes discrimination, bad decisions, or large-scale damage. Reuters said the xAI case has already become part of a bigger fight over whether Washington or the states will set the rules. (usnews.com) So the next phase of United States artificial intelligence policy may look less like one big federal law and more like 50 different traffic systems. One state may require bias checks for hiring tools, another may narrow lawsuits over catastrophic harms, and the companies building the models will test both sets of rules at the same time. (coag.gov) (ilga.gov) (qz.com)
