§ Policy
What policymakers actually need from the AI community now
A short list of things that would make the policy side of the conversation materially better, written from inside the rooms where it happens.
Policymakers working on AI right now are not short of opinions. They are short of operationally useful material.
What is over-supplied
Big-picture futurism, position papers from well-branded organisations, and horizon-scanning reports that identify problems nobody disputes. These are fine. They are rarely what a chief of staff needs on a Wednesday.
What is under-supplied
- Specific, institutionally literate case studies of what has worked and what has not, written in a register that can be circulated internally.
- Frameworks that treat capacity, procurement, and oversight as design problems rather than communications problems.
- Short, credible briefings that a senior official can read in twenty minutes and use in the afternoon.
Where the AI community can help
The most useful contribution a technically literate community can make to policy right now is translation. Not advocacy, not alarm, not product pitches. Translation: taking capability and turning it into language a public institution can reason with. Most of the serious work of the next five years is going to happen through that translation layer. The people who staff it well will have more influence than any single paper.
About the author
Qiqing He works at the intersection of artificial intelligence, workforce transformation, and public-interest institutions. Her work translates technical change into institutional readiness.