The world is being quietly rearranged by people who write very long documents.


The title they went with Administrative Law's Fourth Settlement: AI and the Capability-Accountability Trap Noisy translates that to

Legal scholars propose AI tools to make government decisions auditable


Legal scholars propose new rules for how government agencies use artificial intelligence. These rules would force agencies to explain their AI decisions more clearly.
As government agencies use more complex technology, it becomes harder for courts and Congress to oversee them. This paper suggests a way to make agencies both powerful and transparent. It means agencies could adopt powerful AI tools for things like climate modeling or pandemic response without fear of being shut down for being too opaque.
Watch whether these specific legal ideas, like requiring a "Model and System Dossier" for AI decisions, appear in actual court rulings or legislative drafts.

If you insist
Read the original →