Get Full Government Meeting Transcripts, Videos, & Alerts Forever!
Get email alerts on the AI Verification topic
No spam. Unsubscribe anytime.
Sponsor proposes voluntary AI verification program to build consumer confidence; committee presses on monitoring and liability
Summary
Representative Ty Matthews told the House Technology and Innovation Committee that House Bill 628 would create an optional, market‑based licensing framework for independent verification organizations to certify AI systems; members questioned how certifications would be kept current and what the bill's liability protections mean for consumers.
Representative Ty Matthews presented sponsor testimony for House Bill 628, describing an optional, market‑based approach to AI safety that would license independent verification organizations (IVOs) to evaluate AI systems and provide a transparent mechanism for risk assessment.
"This bill seeks to modernize Ohio's approach to emerging technologies while respecting our state's reputation as a leader in innovation," Matthews said. He said the bill is not intended to impose heavy-handed mandates or intrusive oversight but to provide an optional pathway for developers and public entities to demonstrate adherence to risk‑mitigation practices.
Matthews described a statutory licensing framework for IVOs, a proposed AI safety advisory council to monitor standards, and authority for the attorney general to revoke an IVO's license if it fails to uphold requirements. He used an analogy of a "green check" to explain how verification could signal trustworthiness, and the committee discussed how to keep any designation authentic as models evolve.
Committee members raised multiple concerns: how the state would ensure ongoing monitoring after initial certification, whether monitoring could rely on AI to audit AI, interstate‑commerce limits on state regulation, and whether a voluntary verification regime could advantage large tech firms while burdening small innovators. On civil liability, a member asked how the bill's rebuttable‑presumption language would affect consumers harmed by defective AI products; Matthews said the presumption is intended to balance consumer protection with not overburdening developers and invited drafting suggestions to better protect consumers.
Members and the sponsor agreed the bill's overall framework is a starting point that requires further stakeholder input and drafting to clarify monitoring cadence, enforcement mechanisms, liability consequences, and safeguards against industry capture.
The committee concluded sponsor testimony and adjourned; no committee vote on HB 628 occurred today.
