Missouri missed its chance to hold Big Tech accountable for AI’s dangers | Opinion

The companies building the most advanced artificial intelligence systems describe them in human terms — they “think,” “reason” and “decide.” They deploy AI in roles traditionally held by people: screening job applicants, triaging medical inquiries, managing financial portfolios. And they are racing to build systems that are smarter, faster and more capable than the humans they are designed to serve. But when one of those systems causes harm, the same companies that marketed it as a decision-maker will argue it’s just software — and no human was really responsible.

The Missouri Senate voted 20-10, with bipartisan support, to close that door. Senate Bill 1012, sponsored by state Sen. Joe Nicola, would have made Missouri the first state in the country to move beyond a symbolic declaration that AI is not a person and put an enforceable accountability framework behind it. A House committee voted it down — even after key provisions were stripped under pressure from the White House and industry lobbyists. But the questions the bill raised are not going away, and its framework deserves serious attention from every state’s legislature.

S.B. 1012 asked two questions that will define how AI affects American life for decades: What is AI in the eyes of the law? And who is responsible when it causes harm?

The bill declares that AI systems are products under Missouri law. That single classification pulls the full body of existing product liability doctrine into play — duty of care, design defect, failure to warn — without creating new causes of action or new regulatory structures. The bill does what good legislation should: It applies the rules that already exist to a new technology. And it makes clear that owners and operators bear responsibility when AI causes harm. Any contract that tries to assign fault exclusively to an AI system is void.

Beyond the core framework, the bill addressed specific risks that other states have ignored. Licensed professionals such as doctors, pharmacists and teachers must exercise independent professional judgment and retain final authority over decisions in their practice, even when using AI tools. AI can assist, but it cannot replace the human at the center. Companies cannot use marketing language claiming their AI is “aligned” or “ethically trained” to reduce their legal exposure. Settlement agreements in AI-harm cases cannot include nondisclosure provisions, ensuring that evidence of AI-caused harm is not buried behind confidentiality clauses. And provisions modeled on Sen. Josh Hawley’s federal GUARD Act, which recently passed the Senate Judiciary Committee in a unanimous bipartisan vote, protect children from harmful companion chatbots, requiring suicide and self-harm prevention protocols, AI disclosure and annual safety reporting to the Missouri Department of Mental Health.

Notably, opposition to this bill was backed heavily by technology and trade industry interests that warned the measure could slow innovation or create uncertainty for AI companies. Lawmakers should hear all perspectives. But they should be cautious about allowing the priorities of the companies building these systems to outweigh the interests of the Missourians who will live with the consequences.

Responsible innovation and accountability are not opposing forces. Companies that want to deploy AI responsibly need clear rules as much as consumers do. Without them, responsible developers face the same legal uncertainty as bad actors, and businesses cannot make reliable decisions about how to use AI. SB 1012 would have provided that clarity.

Missouri lawmakers still deserve credit for recognizing the importance of this issue earlier than most states have. This bill might not be moving forward, but the question of whether the law will continue to treat human beings as responsible for the products they create, or allow companies to shift blame onto the technology itself, remains.

Missouri came close to leading the nation on that question. It should not be the last time Missouri lawmakers try.

Riki Parikh is policy director of the Alliance for Secure AI, a 501(c)(3) nonprofit that is working to ensure AI is built with safeguards that serve humanity.