View full title
Establishes understanding artificial intelligence responsibility act requiring developers of covered models to be strictly liable for certain injuries.
In one paragraph
New York's A 8833, the "Understanding Artificial Intelligence Act," would add General Business Law §§ 1510 to 1512. Developers of "covered models" (trained with over 10^26 operations at a cost over $100 million) would be strictly liable, regardless of care exercised, for injuries to non-users caused by model conduct that would be negligence, an intentional tort or a crime if done by an adult human, unless the user or a modifying intermediary intended or reasonably anticipated it. Defamation is excepted. Would take effect immediately. Referred to Science and Technology June 9, 2025 (and January 7, 2026); it has not passed either chamber.
Plain-English summary, written with AI assistance. The official text controls.
What it means for business
None yet: the bill is pending. If enacted, developers of very large AI models would face strict civil liability for qualifying injuries to non-users, subject to the stated defenses.
Key provisions
- "Covered model": an AI model trained using more than 10^26 integer or floating-point operations at a cost exceeding $100 million, as reasonably assessed by the developer
- Except for defamation claims, developers are strictly liable "regardless of the degree of care they exercised" for injuries to a non-user that meet the actual-harm element of negligence
- Liability requires that model conduct factually and proximately caused the injury and was neither intended nor reasonably anticipated by the user or by an intermediary that fine-tuned or modified the model
- For torts requiring a mental state, a rebuttable presumption applies that the AI system had it if a natural person in similar circumstances would have; it is no defense that AI systems cannot have mental states
- Affirmative defenses: the model met the standard of care applicable to humans performing the same function, or the injury resulted from a "capabilities failure" that would not be negligence, an intentional tort or a crime if done by a human
- No penalties or enforcement agency are specified
- Would take effect immediately; introduced June 9, 2025, referred to Science and Technology, with no further action
Latest legislative action
Jan 7, 2026. referred to science and technology
Sponsors
| Name | Role | District |
|---|---|---|
| Alex Bores D | Sponsor | HD-073 |
| Tony Simone D | Sponsor | HD-075 |
Topics and industries
Related topics
Affected industries
This page is general information, not legal advice. Summaries are AI-assisted and may contain errors; read the statute for the controlling text and talk to a lawyer about your situation. Full disclaimer · Report an error