Executive summary: This post argues that the AI governance field suffers from a surplus of abstract research and a shortage of advocacy, leading to a backlog of promising but unused policy ideas (“orphaned policies”); to remedy this, the author recommends that researchers make their work more actionable by drafting concrete policy documents and adopting one of eleven specific, underdeveloped proposals detailed in the post.
Key points:
Advocacy bottleneck: The field has a researcher-to-advocate imbalance (~3:1), meaning many good AI policy ideas lack champions who can bring them to policymakers—creating a large backlog of “orphaned” proposals.
Drafting real policies is tractable and impactful: Researchers should draft actual legislative or governance documents, which are often shorter and more influential than academic papers, and easier for policymakers to act on.
Make white papers concrete and directive: Even without drafting full legislation, researchers can increase their work’s utility by including specific recommendations, estimates, and implementation details.
Legal and funding constraints are navigable: Despite 501(c)(3) limits on lobbying, researchers can still advocate specific policies if framed as nonpartisan analysis with a clear evidentiary basis.
Catalog of eleven underdeveloped ideas: The post outlines eleven “orphaned” policies—ranging from compute monitoring and AI insurance to visa reform and LAWS regulation—each with specific missing components that a researcher could fill in.
Call to action and teaser for next post: The author urges researchers to “adopt” a policy and help develop it for real-world use, and suggests broader institutional reforms (to be discussed in the final post) are needed to shift funding from research to advocacy.
This comment was auto-generated by the EA Forum Team. Feel free to point out issues with this summary by replying to the comment, and contact us if you have feedback.
Executive summary: This post argues that the AI governance field suffers from a surplus of abstract research and a shortage of advocacy, leading to a backlog of promising but unused policy ideas (“orphaned policies”); to remedy this, the author recommends that researchers make their work more actionable by drafting concrete policy documents and adopting one of eleven specific, underdeveloped proposals detailed in the post.
Key points:
Advocacy bottleneck: The field has a researcher-to-advocate imbalance (~3:1), meaning many good AI policy ideas lack champions who can bring them to policymakers—creating a large backlog of “orphaned” proposals.
Drafting real policies is tractable and impactful: Researchers should draft actual legislative or governance documents, which are often shorter and more influential than academic papers, and easier for policymakers to act on.
Make white papers concrete and directive: Even without drafting full legislation, researchers can increase their work’s utility by including specific recommendations, estimates, and implementation details.
Legal and funding constraints are navigable: Despite 501(c)(3) limits on lobbying, researchers can still advocate specific policies if framed as nonpartisan analysis with a clear evidentiary basis.
Catalog of eleven underdeveloped ideas: The post outlines eleven “orphaned” policies—ranging from compute monitoring and AI insurance to visa reform and LAWS regulation—each with specific missing components that a researcher could fill in.
Call to action and teaser for next post: The author urges researchers to “adopt” a policy and help develop it for real-world use, and suggests broader institutional reforms (to be discussed in the final post) are needed to shift funding from research to advocacy.
This comment was auto-generated by the EA Forum Team. Feel free to point out issues with this summary by replying to the comment, and contact us if you have feedback.