Plan A Makes the Case for International Frontier AI Transparency and Pacing

Plan A is a published frontier AI governance scenario that argues international cooperation and full transparency in AI research and development could help prevent a dangerous race toward superintelligence. Developed within the AI Futures Project's AI 2040 materials, it is not an enacted government program or a binding international agreement. It is a recommendation for how policymakers and institutions could deliberately pace the most capable AI systems.

The proposal has gained added relevance amid the Pacing the Frontier open letter, which drew support from more than 1,100 frontier AI employees across companies including OpenAI, Anthropic, Google/DeepMind, and Meta. The letter calls for tools to deliberately pace frontier AI development and for international coordination. Plan A addresses the same broad governance problem, but with a more specific proposed foundation: transparency across AI R&D.

What Plan A proposes

The official AI 2040 Plan A material presents Plan A as a positive scenario and policy recommendation, rather than a prediction. Its central premise is that countries could reach an international deal to avoid competitive pressure pushing AI development ahead of effective safeguards.

Under the scenario, the parties would accept governance guardrails and total research transparency for AI R&D. The intended role of transparency is significant. It would make advanced AI work more visible across borders, potentially reducing uncertainty about rivals' capabilities and making coordinated limits more feasible than they would be in a secretive technological race.

The proposal therefore rests on several linked ideas:

• Frontier AI competition can create incentives to move faster than safety and governance systems can accommodate.
• International coordination is needed because the relevant development and deployment pressures cross national boundaries.
• Transparency is presented as a mechanism for making commitments more credible and for supporting oversight.
• Pacing is framed as an active governance choice, not simply a consequence of technical progress slowing on its own.

That is an ambitious design. It would require participating governments and AI organizations to agree on what qualifies as relevant AI R&D, what information must be disclosed, how disclosures are reviewed, and how noncompliance is handled. The supplied material does not establish that governments have adopted those mechanisms, nor does it provide implementation details for them. Plan A should therefore be understood as a concrete policy direction, not a completed governance framework.

Why the proposal matters for frontier AI governance

The central policy challenge is not merely whether governments should regulate AI. It is whether they can manage incentives that arise when companies and states believe a rival could gain a decisive advantage by moving first. In that setting, domestic rules can be valuable, but they may not resolve international competitive dynamics on their own.

Plan A puts this coordination problem at the center of its argument. A transparency-based arrangement could, in theory, give participants more confidence that others are not secretly advancing capabilities beyond an agreed threshold. That makes it a useful contribution to the policy debate because it links a high-level call to pace development with a proposed institutional principle.

The Pacing the Frontier letter shows that concern about the pace of advanced AI is not limited to outside critics or academic policy discussions. Its signatories include employees from major frontier AI organizations, and reporting by Reuters and Bloomberg has documented the letter and its broad industry representation. OpenAI has also publicly suggested that AI advancement may need to be paced. Together, these developments indicate a live debate over how capability progress should be governed, even though they do not amount to consensus on a single policy model.

There are also practical tradeoffs. Broad transparency may support verification and trust, but the research supplied here does not specify what information would be public, shared only among governments, or protected for security reasons. Nor does it resolve how a framework would account for different national laws, commercial confidentiality, or the risks associated with disclosing sensitive technical knowledge. Those unanswered design questions are central to whether an international transparency proposal could move from scenario to policy.

For organizations evaluating frontier AI governance, the near-term value of Plan A is as a framework for risk management conversations. It directs attention to questions that domestic compliance programs can overlook: how capabilities are monitored across borders, how commitments are verified, and how institutions respond when competitive incentives undermine voluntary restraint. Organizations assessing AI governance and workflow controls can work with Scalevise on AI architecture, automation, and implementation aligned with their operational requirements.

Frequently Asked Questions

What is Plan A in AI governance?

Plan A is an AI Futures Project policy scenario and recommendation that envisions an international deal to pace frontier AI development through governance guardrails and full transparency of AI R&D.

Is Plan A an adopted international AI agreement?

No. The available research describes Plan A as a published scenario and recommendation, not as a legally binding agreement or a policy formally adopted by governments.

How does Plan A relate to the Pacing the Frontier letter?

Both address the need to deliberately pace frontier AI development and support international coordination. Plan A adds a specific proposed mechanism: full transparency of AI research and development.

What does full AI R&D transparency mean in Plan A?

Plan A presents total transparency of AI R&D as a foundation for an international deal. The supplied research does not define detailed disclosure, enforcement, or access rules.

Conclusion

Plan A offers a clear, ambitious answer to a central frontier AI policy problem: how to reduce incentives for an unchecked international race in advanced capabilities. Its transparency-first approach remains a proposal, but its focus on verification, coordination, and deliberate pacing makes it a substantive reference point in the wider governance debate.