The Institute for Progress (IFP) has released a package of 23 specific policy ideas intended to help policymakers begin addressing the risks that arise as artificial intelligence research becomes increasingly automated.
The recommendations are organized into seven thematic categories. IFP characterizes many of the proposals as “low‑regret” steps—measures that are relatively easy to implement and have limited downside. If adopted, the proposals would give governments, especially the United States, additional options as more powerful AI systems are developed. Two explicit aims the IFP highlights are: accelerating the diffusion of AI capabilities by reallocating compute and talent toward inference and new applications, and speeding R&D that makes automated AI research safer either by improving model safety directly or by boosting societal resilience.
The seven categories of proposals
- Provide transparency into automated AI R&D
- Improve state capacity to understand and respond to automated AI R&D
- Develop a risk management strategy for automated AI R&D that accelerates defensive and commercial AI uses
- Accelerate the development of AI verification technology
- Invest in AI resilience
- Extend the US AI lead to give the US more time to manage AI R&D automation risks
- Create option value for international cooperation on managing automated AI R&D risks
These categories cover informational and oversight measures as well as technological and strategic investments. Together, they are designed to expand the set of policy tools available to governments as AI research automation advances.
Why this matters
IFP uses a driving metaphor: today’s global AI development is like driving a car that has only an accelerator pedal. There is little in the way of brakes or advanced telemetry to report speed, engine conditions, or wear. The think tank argues that proposals like these would add the equivalent of pedals and sensing systems to the AI industry’s vehicle, improving the ability to change course or slow down in a crisis.
IFP warns that the fewer options policymakers have for managing the risks associated with increasingly automated AI research, the worse potential outcomes could be. The 23 proposals aim to build a broader, more flexible policy response that includes greater transparency, stronger government capacity, support for verification technologies, investments in societal resilience, and groundwork for international cooperation.
Next steps
The IFP document offers practical ideas for policymakers but is not a binding regulatory plan. Implementation depends on whether governments and other policy actors accept the recommendations and allocate resources to them. Regardless, the paper supplies concrete steps that could be taken if the automation of AI research produces new or heightened risks.
