Taylor Lorenz discusses California SB 1119, a proposal that would require AI services to verify users' identities. She frames the bill as a know-your-customer system for AIKnow your customer is a process for identifying customers and assessing identity or risk before and during a regulated relationship., linking ordinary prompts and account activityPrompt privacy protects the instructions, questions, context, and uploaded content people provide to AI systems from unnecessary access or reuse. to government-issued identity information rather than allowing people to use these tools anonymously or pseudonymously.
Lorenz argues that the consequences extend beyond inconvenience. Whistleblowers, vulnerable communities, researchers and ordinary users may avoid sensitive questionsA chilling effect occurs when people avoid lawful speech, inquiry, or participation because they fear monitoring, exposure, punishment, or other consequences. if every interaction can be traced back to them. Centralized identity recordsA centralized identity database stores identity records under one controlling system or authority instead of distributing or minimizing them. could also become targets for breaches, subpoenas or cross-platform tracking.
The interview places the proposal inside a wider policy conflict involving child safety, free expression, censorship and national competition. Lorenz says lawmakers should address genuine harms without making universal identity collectionData minimization limits information collection, use, access, and retention to what is necessary for a defined purpose. the default, particularly when enforcement mechanisms can outlast the political moment that created them.
Watch on YouTube
