How to Tax AI, by @petercoy @dealbook bit.ly/4eFKFRD Jeremy Bearer-Friend (@gwlaw) & @SPolcz (@UCDavisLaw)

This is another reason why I've recently become human-pilled for law. Law is where we operationalize our values as a basis for resolving disputes. It's tempting to turn that over, even in part, to AI—it would be more efficient and cheaper. But we don't need to do that. Recently there's been a lot of talk, led by people working in @AnthropicAI but joined by a lot of legal scholars and others, of how AI-assisted legal research and filings are already showing signs of potentially overwhelming our already stretched legal system. And, in turn, how this motivates the use of AI in supporting and, in the case of arbitration, even replacing human decision-making. But I don't understand why we need to go there? We're already headed there with broader education, but especially if AI delivers on its promise, we are shortly going to be in a situation where there's a surplus rather than shortage of knowledge workers. In that case, we have an opportunity to greatly expand the population of people who can be counsellors, advocates, jurists, and neutral arbiters and mediators, and other participants in dispute resolution. I've been wondering why more people aren't thinking this way. We can be look to, for example, the response to the growth of AI in creative industries where more and more people who previously lacked the resources to do expensive CGI and viz effects can now have access to that more cheaply and to create things they couldn't do otherwise.

I don’t understand why this consideration of “being kind” to a model even enters the calculus. Are you “kind” to a calculator? Are you “kind” to your remote control? I’m sorry but allowing emotion to enter your mindset when interacting with a model is a symptom of AI psychosis. And to be clear, I’m not saying “cruelty” is good either. In fact, it is just as bad because, in both instances, you are attributing human emotion to an inanimate object. A model is a thing. You simply use it. There is no “kind” or “cruel;” there is “productive” and “unproductive”—that’s it. Anything more is allowing your mind to get twisted into believing a weird fairy tale.

I have a bold suggestion for America’s next AI Czar: Put the people in charge of AI’s future.

Are you ready for a radical idea? Billionaires should not buy our elections. One person, one vote.

The President is choosing an AI Czar, and I'm throwing my hat in the ring (dark horse though I may be). From developing Google Drive to founding Asana and One Project, I've spent a career helping people collaborate on hard problems. Collaborating to steer the future of AI is our hardest problem yet. AI will supercharge whatever goals we give it. It could cure diseases, create industries, and make life more affordable for everyone. But if we aim its enormous power at the wrong goals, even unintentionally, the results could be catastrophic. I saw it happen after I helped create Facebook's Like button. This isn't a partisan issue. We don't have to choose between the President's vision of "building America's Golden Age" and protecting human wellbeing. They're one and the same. As AI Czar, I would: 1. Let Americans decide what to do with the money AI generates. President Trump wants Americans to share in AI's success, calling it "a partnership with the American public." He's right. A share of AI's wealth should go to the public: we, the people, whose cultural output is the resource these companies mine, and whose lives they'll disrupt. But if AI does all the work, and only a few people own it, a dividend check alone leaves the rest of us with no leverage at all. And handing the money to political appointees ties it to politics. Instead, I'd push for participatory budgeting, where Americans decide together how to invest it. In Columbus, Ohio, residents are practicing it to decide how to spend $9 million. 2. Give Americans a governing stake in major AI companies. Senator Sanders has proposed the U.S. take a governing stake in the labs. I'd go one step further and put it in the hands of the American people, through a citizens' assembly on AI, not another Washington commission. 3. Lead the world toward international cooperation, like we did with nuclear safety. The President has said, "Whoever wins AI, wins." But if AI puts humanity in danger, everyone loses. Just as America pioneered nuclear safety protocols and then led the creation of the International Atomic Energy Agency, we can lead on democratic control of AI, with citizens' assemblies of everyday Americans, chosen by lottery and supported by experts, proposing the goals AI must serve and the guardrails it must honor. Washington would have to adopt each proposal or publicly explain why not. Assemblies are already shaping AI policy in the UK, Belgium, and Taiwan. Ours would also lend democratic legitimacy to U.S. proposals in talks like the ones Secretary Bessent recently began with China. President Trump is optimistic about the future of AI, and I want to be too. As AI Czar, I would do what I do best: help people collaborate. Together, we can prevent the worst and realize the full promise of AI: a stronger, healthier, more prosperous America, and world. If you agree that Americans deserve a say in AI's future, please share this, and share your ideas (or feedback on mine) in the comments.