What Should Be Done

2026-03-02 · Source: Hyperdimensional · Field: Technology & Digital — Artificial Intelligence & Machine Learning, AI Governance & Policy · Depth: Intermediate, long

Summary

The US government's Executive Order on Cyber and AI has created a de facto involuntary licensing regime for frontier AI models, exemplified by restrictions on Anthropic's Fable and OpenAI's GPT 5.6. This policy, intended to address catastrophic risks, lacks clear safety standards, leaving labs without defined release requirements. This uncertainty impedes labs' ability to recoup significant training costs and threatens the US AI infrastructure buildout, which relies on a global market. The author proposes a solution for the next 6-18 months: federalize state laws requiring labs to disclose safety frameworks and establish independent verification organizations (IVOs). These IVOs would audit labs' adherence to their safety plans and internal governance, fostering real-world safety experience and best practices. This approach, which could see government certification of auditors, aims to regulate the "frontier lab as an entity" rather than individual, rapidly evolving models, freeing government resources for other critical areas like biosecurity.

Key takeaway

For policy makers developing AI governance frameworks, you should prioritize establishing clear, auditable safety standards. Relying solely on government pre-approval without defined criteria risks stifling innovation and jeopardizing US AI infrastructure. Instead, certify independent verification organizations to assess labs' adherence to their evolving safety frameworks. This fosters real-world experience and best practices. This approach frees government resources for critical infrastructure and biosecurity, ensuring robust, adaptable oversight.

Key insights

Government's de facto AI licensing without clear standards stifles innovation and infrastructure, necessitating independent verification.

Principles

• AI safety standards require real-world experience and iteration.
• Broad AI diffusion is essential for economic and defensive benefits.
• Regulating frontier labs as entities is more robust than regulating models.

Method

Federalize state laws mandating public disclosure of lab safety frameworks. Establish independent verification organizations (IVOs) to audit labs' adherence to these frameworks and internal governance, potentially with government certification.

In practice

• Consult public safety frameworks from leading AI labs.
• Support legislative efforts to federalize AI safety disclosure laws.
• Investigate models for independent technical auditing bodies.

Topics

• Frontier AI Models
• AI Regulation
• AI Safety Standards
• Independent Verification Organizations
• AI Governance
• US AI Policy

Best for: CTO, VP of Engineering/Data, Executive, Policy Maker, Director of AI/ML, Legal Professional

Related on AIssential

• The best AI bill yet may not get far — The Great American AI Act represents a comprehensive, yet politically challenged, federal effort to establish AI safety standards through mandatory audits and a contentious preemption framework.
• OpenAI unveils security framework, citing emerging regulations — AI providers are proactively establishing governance frameworks to meet diverse, rapidly evolving state and international regulatory demands.
• What Washington must do — Opaque, rushed AI regulation risks corruption, stifles innovation, and undermines national leadership, necessitating transparent, principled governance.
• Opinion | AI Regulation Needs a ‘Bank Examiner’ — AI regulation is shifting towards a "bank examiner" model, emphasizing expert supervision over product-safety laws.
• The Once And Future Fable #5 — The "Mythos Moment" reveals an ad hoc, reactive US AI policy struggling with frontier model control, legal frameworks, and open-source risks.
• Spooked by Mythos, Trump suddenly realized AI safety testing might be good — Government AI oversight is shifting towards pre- and post-release safety evaluations, but faces challenges in defining standards and ensuring independence.

See Counsel's argued verdicts on the open AI decisions leaders are weighing →

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Hyperdimensional.