Bill > S2081

This bill establishes a conditional immunity framework for artificial intelligence (AI) developers when their AI products are used by licensed professionals in their work. To qualify for this legal protection, AI developers must publicly release and continuously maintain a comprehensive "model card" that details the AI system's training data, performance metrics, intended uses, and limitations, as well as a "model specification" that explains the system's core behavioral instructions. Developers must also provide clear documentation about the AI product's known limitations and appropriate use cases. The immunity applies only when the AI is used by a "learned professional" (such as a licensed healthcare provider, lawyer, or financial expert) and does not cover cases of recklessness or willful misconduct. Developers must update their documentation within 30 days of deploying a new version or discovering significant system flaws, and failure to do so can void their immunity. The bill aims to balance innovation in AI technology with responsible development by creating a transparent framework that helps professionals understand and appropriately use AI tools while protecting developers from broad liability. The legislation will take effect on December 1, 2025, and apply to acts occurring on or after that date, with explicit provisions preventing its use to shield developers from fraud or deliberate misrepresentation.

Transportation and Infrastructure

https://www.congress.gov/bill/119th-congress/senate-bill/2081/all-info