The finding
In the absence of any federal law governing AI systems, states have taken up chatbot safety. NPR reported on Thursday that in at least ten states, chatbot bills introduced this year carry similar language — and that several of the shared provisions would exempt the most widely used chatbots on the market.
NPR’s account is based on interviews with lawmakers, policy analysts, lawyers and technology safety advocates, and a review of bill text and state lobbying disclosures. Several of the bills have already been signed into law.
The pattern itself is unusual. Successful state legislation normally spreads slowly, one state copying another over years. Laura Marquez-Garrett, an attorney at the Social Media Victims Law Center, told NPR that seeing the same language appear simultaneously in ten states is not how this usually works.
The shared carve-outs
Three exemptions recur. Bills in at least eight states, including Idaho, Nebraska, New Jersey and Oklahoma, exempt “a feature within another software application, web interface, or computer program” — which policy analysts told NPR could cover Meta AI and Grok, embedded in X.
Five states’ bills carry identical language exempting any chatbot that “functions as a speaker and voice command interface or voice-activated virtual assistant for a consumer electronic device”, which would apply to Amazon’s Alexa. Modified versions appear in a further ten states’ bills.
A third pattern limits a bill’s reach to minor “account holders”, which would exempt chatbots a person can use without registering — ChatGPT and Google Gemini among them.
Colorado as the worked example
Colorado’s House Bill 26-1263 was introduced in February and written in collaboration with Healthier Colorado, a state health policy non-profit. A person with knowledge of the process, speaking anonymously to NPR for fear of professional reprisals, said Healthier Colorado worked with Google on the bill.
The enacted law excludes twelve categories of chatbot, among them products primarily designed for developers or researchers, commerce and customer-service assistants, chatbots housed within search engines, and products marketed for business productivity. Marjorie Connolly of the Tech Oversight Project told NPR the carve-outs cover “developer tools, video games, voice assistants, and chatbots built into social media platforms”, and that “parents are being asked to accept a chatbot law that exempts chatbots”. On her reading, ChatGPT, Microsoft’s Copilot and Anthropic’s Claude could each argue they fall outside it; Gemini could qualify under the search-engine exemption.
State Rep. Sean Camacho, a Democrat and one of the bill’s sponsors, defended it. “We could not pass a bill that would make Colorado an outlier,” he told NPR, and added: “You can’t legislate tech companies unless you have their input.” He said he worked with Healthier Colorado and other stakeholders including Google, and that Governor Jared Polis asked him to make the bill resemble those in other states. Healthier Colorado’s spokesperson Kate Morr said the bill “is not an industry bill in disguise” and described the law as “a floor we intend to keep building on, not a ceiling”.
The lobbying record
Google lobbyists registered in support of chatbot bills in Iowa, Colorado, Nebraska and Arizona, according to state disclosure records NPR reviewed. Hawaii state Rep. Trish La Chica said Google lobbyists approached her to sponsor a chatbot safety bill written by the company. A Republican Colorado lawmaker, speaking anonymously, said a Google lobbyist approached them to sponsor a bill the lobbyist was “shopping”, and declined: “Google has gone on the offensive when it comes to writing legislation that specifically carves themselves out of situations.”
In a statement to NPR, Google said it “is committed to supporting thoughtful, effective AI legislation that protects consumers while fostering innovation” and that it “actively collaborate[s] with lawmakers, industry peers, and community advocates to help shape policy frameworks that promote safe, reliable, and beneficial AI tools”.
Why this is being legislated at all
NPR’s review of court records found at least 75 lawsuits filed in federal and state courts against AI developers over alleged harms from chatbots, many involving children. Character.AI settled multiple such suits in January. Most of this year’s state bills are modelled on California’s SB 243, which passed in 2025 and which consumer advocates regard as industry-friendly.
The legislative sessions that produced these bills are over in most states. The open question is whether the next round is written the same way.
