On Tuesday, Reps. Lori Trahan (D-MA) and Jay Obernolte (R-CA) introduced the FRONTIER Act, legislation that would establish a federal framework for governing advanced AI models, including independent third-party assurance, mandatory safety frameworks, and enforcement authorities. The bill includes language that would preempt state safeguards, a controversial proposal that stirred opposition when a discussion draft of the bill was floated last month. The legislation also fails to provide any avenue for whistleblowers to report a violation of the proposed regime.
“We are pleased to see that the sponsors of the FRONTIER Act have addressed some of the critical concerns raised by public interest groups following the publication of their discussion draft last month, but key concerns remain,” said ARI President Brad Carson. “The bill’s preemption clause continues to leave a serious hole in regulatory oversight of AI models. For example, the bill sets up a federal private assurance market to oversee companies’ safety efforts, but that market will take years to set up. Meanwhile, the FRONTIER Act’s preemption of state laws would be effective immediately. Drafting issues also create the risk that what is intended to narrowly preempt state laws related to catastrophic risk may be applied more broadly to laws relating to safety issues created by AI for patients and consumers. There’s also no sunset of the bill’s preemption provision, permanently tying the hands of states on AI laws.”
“Additionally, as written, the FRONTIER Act leaves the decision to release AI models that pose a serious threat in the hands of the developers. The current bill leaves AI companies to perform a cost-benefit analysis on whether deployment of a high-risk model is acceptable. Whether that analysis and determination should sit with AI labs that have a strong incentive to release risky AI systems is worthy of serious debate both within Congress and amongst the public. To its credit, the current bill proposes a mechanism by which the government can halt deployment, but leaving the initial deployment decision of risky models in the hands of AI labs should not be taken lightly. We appreciate the attention of Reps. Trahan and Obernolte to the need for federal AI guardrails and look forward to their continued work on this legislation.”
###
Americans for Responsible Innovation (ARI) is a nonprofit organization dedicated to policy advocacy in the public interest, focused on emerging technologies like artificial intelligence (AI). Learn more at ARI.us.
