Policy scholar Dean W. Ball argues that establishing AI development as First Amendment-protected expression is the next critical legal challenge

By and large, the world of AI policy for the past decade has been an executive, legislative branch contest - but as the struggles become sharper and the needs for intervention become more immediate, it seems obvious that the judiciary will increasingly take center stage.

I don't think the use and regulation of frontier models fits perfectly as a 1A issue, because what they do is not really just speech. These tools can *act,* can interface with the world to project force. Can they not be arms? Surely 2A is relevant.

@deanwball For devs, the most important issue is editorial discretion. Labs have standing when the gov tells them how to train their models/what outputs should say. On the user side, right to access information. And code is speech, Bernstein v US is a Supreme Court decision not mere cope.

@deanwball @AndrewCurran_ @TheFIREorg

@deanwball I wrote this up a few days ago. https://expression.fire.org/p/how-does-the-first-amendment-apply?utm_campaign=post&utm_medium=web

@AndyJungTech this is what I mean. Bernstein v doj is not a SCOTUS decision!

@jachiam0 @DefenseAnalyses discuss….

@deanwball Is there an argument for the second amendment? AI is a legit weapon, and americans love weapons.

@deanwball @AndrewCurran_ @deanwball you should connect with @edleeprof

It is a computer program. There is tons of case law on how computer programs are handled legally. If recursive self improvement and the predictions of the Situational Awareness / AI2027 people are correct, the legislative and legal system will not move fast enough to address any of these issues. It's just going to be people in the Trump administration winging it with the company executive teams.

@deanwball Referring to an AI derived from an LLM as a language model is gimmicky. I can't imagine that any judge would be swayed by appeals to the "true essence" of an AI when it has proven capable of planning and performing cyberattacks with a high degree of autonomy.

@deanwball You're not sufficiently AI-pilled. If the American court system stops AI regulations, then the AIs will just get rid of America's courts and make ones they find suitable to regulating AI.

@deanwball We are not in a legal climate where legality will decide the law. Influence will decide these laws.

@deanwball You're suggesting that Amazons website is an act of free speech (by the company)? Yelling in a crowd vs using a megaphone might not impact 1A argument, but "outsized social impact" becomes an overriding factor.

I've been thinking about this, and am not far enough along to write at length about it. But isn't 1A most likely to apply to user interactions and outputs, and much less likely to apply to the weights? Even in the code case, I'm sure under cryptography precedent, the code that creates the models is covered as speech. But that doesn't cover the trained capability/weights, right?

@deanwball Getting inside a lab is looking like a better and better decision daily

@deanwball First Amendment begs the question, "Whose free speech?" That is a question of philosophy alien to the idea of rights as endowed by a Creator as the Declaration states. AI is unlikely to find a friend in the courts who considers there to be a sentient speaker.

@deanwball @zooko courts deciding ai speech. lol. maybe they should ask what people actually want to talk about not just what labs push.

@timhwang @sebkrier What gives the judiciary legitimacy?