Senators Richard Blumenthal (D., Conn.) and Josh Hawley (R., Mo.) have proposed a new legislative framework to regulate AI. This framework includes expansive powers for an oversight body. A summary of the role and powers of the proposed oversight body is below:
- Oversee a federal licensing process for AI. Licensing requires registration of AI model information alongside risk management, pre-deployment testing, data governance, and adverse incident reporting programs.
- The oversight body would be able to audit companies seeking licenses and monitor/report on AI's technological and economic impacts.
- The oversight body would cooperate with other enforcement entities and have concurrent enforcement authority with state attorneys general.
- Development of conflict-of-interest rules by the oversight body to address "capture" and "revolving door" concerns.
- The oversight body would enforce legal accountability when AI models cause harm, with a private right of action in case of privacy breaches or civil rights violations.
- Establishment of a public database by the oversight body for easy access to AI model/system information.
- The oversight body would play a role in ensuring companies deploying high-risk AI implement specific "safety brakes."
- The oversight body would advocate for consumer control over personal data in AI systems and set strict boundaries for generative AI involving children.
This is just a proposal and is not yet law. It is not clear what the odds are that this framework becomes enforceable law. Companies developing and using AI may consider aligning their AI practices with the proposed framework to prepare more readily for AI regulation on the horizon.