Major artificial intelligence (AI) developers would have to disclose whether and how extensively they use their own technology to develop new models under a policy proposal aimed at giving federal officials greater visibility into a potentially accelerating cycle of AI advancement.
The framework from Americans for Responsible Innovation (ARI), a nonprofit comprised of academics, former congressional staffers and government officials, would require quarterly reports to a new Commerce Department office. Significant changes in research practices would trigger notification within four business days, according to Nextgov/FCW, which first reported the proposal.
The central concern is that AI could increasingly perform the research needed to build more capable successors, shortening development timelines and leaving government officials and safety researchers less time to evaluate emerging risks. ARI wants reporting obligations to capture that progression before a security incident exposes a problem.
“Washington’s visibility into that shift should not depend on what companies choose to announce,” the group wrote in a blog post describing its framework.
The proposal comes as a Trump administration task force engages with industry and outside groups to assess AI risks and shape the government’s relationship with developers. President Donald Trump tapped Director of National Intelligence Jay Clayton this month to lead the “Super Intelligence Force,” established by an executive order signed in September. “Super intelligence” is the Trump administration’s preferred descriptor for artificial intelligence.
Following its first formal meeting, the task force assigned Pentagon technology official Emil Michael to develop protocols for laboratories to report breaches, hacks and other cybersecurity problems. It also commissioned a study covering cybersecurity, biological threats and the ability to maintain control over AI systems.
ARI’s proposal would extend government visibility beyond incident reporting. Companies would disclose how much research AI performs and how closely humans supervise that work, even when nothing has gone wrong.
We’d love to be your preferred source for news.
Please add us to your preferred sources list so our news, data and interviews show up in your feed. Thanks!
Developing AI involves writing software, preparing training data, running experiments and deciding which results deserve further investigation. Automating those activities could allow developers to produce stronger systems faster. A change such as fully automating a research cycle could trigger the proposed four-business-day notification requirement.
For oversight purposes, the distinction matters. Incident reports describe failures that have already occurred, while research disclosures could help officials identify changes in development speed and human supervision that warrant closer examination.
The framework would give the Commerce Department authority to seek additional information and require more frequent reporting. Knowingly false or incomplete filings could be referred to the Justice Department. ARI also proposed a whistleblower program, adding another channel for information about developers’ practices.
Its proposed safety standards would address hacking capabilities, large-scale influence operations and chemical, biological, radiological, nuclear and explosive threats. The breadth reflects concerns that increasingly capable systems could create risks across several areas.
The blueprint would draw on technical analysis from the Commerce Department’s existing National Institute of Standards and Technology (NIST) center, known as the Center for Advancing Innovation and Standards for Super Intelligence. That center already works with developers under voluntary agreements to evaluate advanced models, including their hacking capabilities.
Under ARI’s approach, the center’s analysis would inform safety standards and help determine which models should fall under federal oversight. The proposal leaves unresolved whether regulatory responsibilities would remain within Commerce or be divided among agencies.
The administration has emphasized company-led safeguards while resisting a broad new regulatory regime. A September industry accord calls for internal controls, outside audits and independent oversight committees, and Clayton’s task force is working to implement it.
ARI’s framework would move toward mandatory reporting about the development process itself. For companies, the practical implication would be a need to document AI’s research role, human supervision and significant changes as development proceeds. Whether the administration embraces that approach will help determine how much federal oversight depends on voluntary industry cooperation.
Add as Preferred Source
