To be clear, there is a lot of uncertainty as to how longstanding rules of administrative law will apply to these new technologies. We get into that too. Read up! governingforimpact.org/wp-content/u...

But administrative law will let the sunlight in on agencies' AI usage. And that should go some way toward pushing agencies away from unscrupulous and irresponsible uses of new technology toward more defensible practices.

These rules of transparency shouldn't necessarily limit an agency's ability to use AI—after all, it's easy to imagine many ways AI systems could, if used responsibly, make federal rulemaking faster, more effective, and more open.

First, agencies will often need to DISCLOSE when they have used AI to power a rulemaking. Second, they must EXPLAIN how they used AI, and why they think AI has produced reliable results in particular cases.

The Administrative Procedure Act broadly requires agencies to be transparent about how they make rules. They must make public the data and assumptions on which they rely and explain themselves. We think those rules extend to AI usage.

Topics covered:

> FINAL AGENCY ACTION
> NOTICE-AND-COMMENT RULEMAKING
> ARBITRARY-AND-CAPRICIOUS REVIEW
> PRETEXTUAL AGENCY ACTION
> NON-APA REVIEW OF EXECUTIVE ACTION
> REMEDIES
> AGENCY NONENFORCEMENT
> HOW TO WRITE A COMMENT

An administrative law casebook you can use!

Opinion here: storage.courtlistener.com/recap/gov.us...

Judge Liman, in granting a preliminary injunction against the "cancelation" of congestion pricing, agrees with this reading:

Count yourself lucky! Dog images are treated as a currency in heterosexual culture.

Do go on!

Filing: storage.courtlistener.com/recap/gov.us...