Show Notes
Artificial intelligence is usually sold as productivity: faster research, faster drafting, faster answers. In this episode of New Law Order, Yale Law professor Scott Shapiro explains that the real story is power. If the law is a system for coordinating human behavior at scale, then tools that can interpret rules, test edge cases, and generate persuasive legal analysis may change who can navigate the system—and who gets steamrolled by it.
The conversation also explores a darker symmetry: the same tools that help people comply with legal obligations can help sophisticated actors evade them. Shapiro sketches the emerging role of artificial intelligence as both compliance engine and exploitation engine—and what that means for the balance of justice when the best “legal hacker” is a machine.
Shapiro draws a crucial distinction between generative models and rule-based systems. Generative tools can sound authoritative while being intentionally or accidentally wrong, which can present dangers. Rule-based approaches, by contrast, raise a different question: what happens when the logic is correct but the reasoning is no longer human-followable? At that point, “legal reasoning” starts to look more like an appeal to authority—raising uncomfortable issues about legitimacy, transparency, and trust.
In this conversation:
Law as a Social Technology: why legal systems exist and what they are built to do Generative Models versus Rule-Based Reasoning: reliability, hallucinations, and explainability in legal work When Legal Reasoning Stops Being Human: what legitimacy requires when the logic becomes unreadable Artificial Intelligence as Exploitation: using models to find loopholes, stress-test regulations, and “break” contracts
Guest: Scott Shapiro is a Yale Law School professor and legal philosopher whose work explores the structure of legal systems and how technology changes the way law is interpreted and applied.
For more conversations with the titans changing the legal business, subscribe to New Law Order. For a lawyer Continuing Legal Education version of this interview and other legal insights, visit www.talksonlaw.com. For questions or comments about the show email us at [email protected].
More from New Law Order
Big Law has spent decades and a fortune on diversity, and the numbers barely moved at the partnership level. The usual explanation is that somewhere in the building (or somewhere in your head) there's racism and misogyny. David Wilkins, Harvard Law professor and possibly the most important scholar on the legal profession, has a less satisfying answer: you don't need bigots to get this result. You just need busy partners.
In this episode of New Law Order, Joel Cohen and Yale Law Professor John Morley speak with CJ Mahoney, Chief Legal Officer of Meta, about one of the hardest legal questions facing modern technology companies: when a product becomes deeply woven into daily life, how should the law think about harm, responsibility, and regulation? Drawing on a career that has taken him from Williams & Connolly to the Office of the U.S. Trade Representative, Microsoft, and now Meta, Mahoney offers an unusually candid look at what it means to serve as the top lawyer inside a company operating at the center of law, politics, and technological change.
Richard Sander, UCLA law professor and economist, has spent decades advancing one of the most controversial arguments in legal education: that large racial preferences in law school admissions may frequently harm the students they are intended to help. In this episode of New Law Order, Joel Cohen and Yale Law Professor John Morley speak with Sander about affirmative action, the Supreme Court’s decision in Students for Fair Admissions, and what new data may reveal about the future of law school admissions.
Michael A. Gerstenzang, the longtime managing partner and now senior partner of Cleary Gottlieb Steen & Hamilton, explains how he positioned one of the world's most established law firms to lead in AI adoption — not by outsourcing innovation, but by building it from the inside. The conversation traces a series of structural bets that began before generative AI entered the mainstream, starting with ClearyX, a wholly owned technology subsidiary launched in 2021 to reinvent M&A due diligence. Gerstenzang describes the venture's founding provocation — imagine the business that puts Cleary Gottlieb out of business, and build it — and how A/B testing against traditional associate-led diligence demonstrated equal or superior quality at roughly half the cost.
This week, we’re sharing a timely episode from a related TalksOnLaw podcast, AI Lawyer. In this episode, host Joel Cohen is joined by Judge Paul Grimm for a conversation about one of the most unsettling problems facing the legal system: what happens when audio, video, and other forms of evidence can be generated, manipulated, or challenged by artificial intelligence?