When an AI system a company sold gives dangerous advice, invents a defamatory "fact" about a real person, or hands an attacker the keys to the system it was trusted with, the company's answer is a terms-of-service page you clicked past and a claim that nobody could have seen it coming. Every other industry that ships a product into people's lives (cars, drugs, ladders, baby formula) answers for foreseeable harm in a courtroom. That isn't hostility to the technology; it's the ordinary price of selling something, and it's the only mechanism that has ever reliably gotten a company to fix a defect before the lawsuit instead of after. Pass a federal AI liability standard: no blanket immunity, no forced arbitration, and a clear path to court when a company's system causes real harm.
When an AI system a company sold gives dangerous advice, invents a defamatory "fact" about a real person, or hands an attacker the keys to the system it was trusted with, the company's answer is a terms-of-service page you clicked past and a claim that nobody could have seen it coming. Every other industry that ships a product into people's lives (cars, drugs, ladders, baby formula) answers for foreseeable harm in a courtroom. That isn't hostility to the technology; it's the ordinary price of selling something, and it's the only mechanism that has ever reliably gotten a company to fix a defect before the lawsuit instead of after. Pass a federal AI liability standard: no blanket immunity, no forced arbitration, and a clear path to court when a company's system causes real harm.
