“A bill to establish immunity from civil liability for certain artificial intelligence developers, and for other purposes.”
No CRS summary available for this bill.
This section states seven congressional findings concerning the rapid advancement and deployment of artificial intelligence systems in professional services (e.g., healthcare, law, finance), the lack of clarity on liability for AI errors, the need for transparency on AI capabilities and limitations, and the importance of standards and frameworks to promote responsible innovation and allocate responsibility among developers and users.
This section establishes definitions for terms used in the Act, including (1) artificial intelligence, which has the meaning given in section 5002 of the National Artificial Intelligence Initiative Act of 2020 (15 U.S.C. 9401); (2) client, as a person engaging and relying on a learned professional's services under professional standards; (3) developer, as a person creating, controlling, or marketing an artificial intelligence product; (4) error, as a false, misleading, or incomplete output or failure to perform a represented function; (5) learned professional, as a licensed individual exercising independent judgment bound by professional standards; (6) model card, as a public document detailing an artificial intelligence product's training data, performance, uses, limitations, and risks; and (7) model specification, as developer-supplied instructions (including system prompts and training guidance) establishing an artificial intelligence product's behavior and outputs.
This section establishes conditional civil immunity for developers of artificial intelligence (AI) products when the product is used by a learned professional (i.e., a licensed or certified practitioner such as a physician or attorney) in providing professional services to a client, if the developer (1) publicly releases and continuously maintains a model card and model specification for the product (with redactions permitted only for trade secrets unrelated to safety, justified in writing) and (2) provides clear documentation to learned professionals on the product's known limitations, failure modes, and appropriate uses. The immunity excludes recklessness or willful misconduct by the developer and lapses if the developer fails to update the model card, specification, and documentation within 30 days of deploying a new version or discovering a new material failure mode that proximately causes subsequent harm. This section preempts state law claims against immune developers for AI errors in this context but does not affect claims based on fraud, knowing misrepresentation, or use outside professional services.
This section preserves immunities from civil liability established by federal or state law or available at common law that are unrelated to the immunity established under section 4(a).