You are watching a webinar preview. To view the full webinar, log in to your Justia Connect account or sign up for free.

AI and Legal Ethics: Navigating New York’s Evolving Framework

In-Person Location: 25 W 39th Street, 9th Floor New York, NY 10018

This program argues that the profession’s compliance-model response to AI risk (policies, training, exhortation to verify) is insufficient, and that most AI ethics problems lawyers face are better understood as delegation and supervision problems. Drawing on the Sullivan & Cromwell filing, 22 NYCRR Part 161, United States v. Heppner, and NYC Bar Formal Opinion 2025-6, the session provides a practical framework for distinguishing tasks lawyers can safely delegate to AI from judgments they cannot, tiered verification guidance matched to different categories of AI-assisted work, and concrete guidance on the emerging obligation to counsel clients about AI-related privilege and confidentiality risks.

Duration of this webinar: 60 minutes

Originally broadcast: July 24, 2026 12:30 PM PT

Webinar Highlights

This webinar is divided into section summaries, which you can scan for key points and then dive into the sections that interest you the most.

00:00 AI in Legal Practice: Case Study and Challenges

David Kemp, the speaker, begins by discussing a case study involving a prominent law firm's bankruptcy petition riddled with errors due to AI use. He highlights the importance of delegating tasks to AI while retaining judgment, and the ongoing debate about AI disclosure in New York courts. Kemp notes that many AI-related errors go unnoticed in transactional work, emphasizing the need for thorough verification processes. He stresses the importance of structural changes to ensure AI outputs are verified before submission, drawing parallels to medical practice checks.

15:29 Delegating Tasks to AI: Risks and Strategies

Kemp discusses the risks of delegating judgment to AI, emphasizing the need for human oversight in legal tasks. He explains that AI lacks context and cannot engage in legal reasoning, making it unsuitable for tasks requiring discretion. Kemp advises on how to phrase AI prompts to retain human judgment, using specific examples to illustrate effective strategies. He suggests using AI as a thought partner for generating options, rather than relying on it for evaluative conclusions.

26:33 Disclosure Policies and Client Communication

Kemp outlines New York's policy on AI disclosure, noting that it does not require attorneys to disclose AI use in court submissions. He contrasts this with other states' varying disclosure requirements, highlighting the benefits of disclosure in supervisory contexts. Kemp discusses the implications of clients using AI, emphasizing the need for clear communication about AI's risks. He advises on maintaining secure communication channels and providing alternatives to prevent clients from relying on AI for legal advice.

41:42 Unchecked AI Use and Verification Strategies

Kemp highlights the risks of unchecked AI use in legal documents, particularly in transactional and regulatory contexts. He suggests structural changes and habits to ensure AI outputs are verified, such as using adversarial prompting and maintaining verification logs. He draws parallels to information security practices, advocating for proactive measures to prevent AI errors from perpetuating. Kemp concludes by emphasizing the importance of AI competence over mere compliance, advocating for task delegation while retaining judgment. He highlights the role of disclosure in efficient supervision and advises on counseling clients about AI use.

Please note this AI-generated summary provides a general overview of the webinar but may not capture all details, nuances, or the exact words of the speaker. For complete accuracy, please refer to the original webinar recording.

Speaker

David Kemp AI & Innovation Training Manager
Ropes & Gray LLP

David S. Kemp is a part of the AI innovation team at the global law firm Ropes & Gray LLP, and the managing editor of Justia’s Verdict and Oyez. He has taught at Rutgers Law School, the University of California, Berkeley School of Law, and UC Law SF (formerly Hastings College of the Law). His areas of teaching include lawyering skills, professional ethics, and generative AI.