The AI Systems Transparency Act would require the Federal Trade Commission (FTC) to set disclosure rules for developers of certain widely used generative AI models and for large services that use covered AI systems. The rules would require public information about model capabilities, safety testing, privacy practices, safeguards for minors and other vulnerable users, and changes in model use or behavior. Disclosures would include a version for consumers and a more detailed version for independent researchers. The bill would allow some security-sensitive details and trade secrets to be withheld. The FTC would enforce the requirements.
If the bill became law, you could find information about covered AI systems on developers’ websites, including how the systems are tested and what data and safety practices are used. The bill does not require every AI system or service to disclose this information; it sets user thresholds and allows the FTC to designate systems that pose significant safety risks.
The bill provides no cost estimates or specific funding amounts. It would require the FTC to make rules and provide technical assistance, and covered companies would have disclosure duties.
No publicly available information.
No publicly available information.