Does attorney-client privilege exist when you use ChatGPT or Claude? Should it?

Jacob Robinson sits down with Mike Katz, Partner at Manatt Phelps & Phillips, to examine whether attorney-client privilege, work product or any analogous protection applies when people ask AI chatbots legal questions.

Timestamps:

➡️ 0:00 — Background

➡️ 1:47 — What is attorney-client privilege?

➡️ 2:44 — Policy reasons for narrowing privilege

➡️ 3:30 — The Upjohn case (1981)

➡️ 4:34 — Privilege vs. work product doctrine

➡️ 5:17 — Three elements to establish privilege

➡️ 7:23 — Consumer AI terms of service and confidentiality

➡️ 8:09 — How you lose privilege

➡️ 11:30 — War stories

➡️ 15:39 — Vibe lawyering

➡️ 19:09 — Could Anthropic, OpenAI be liable?

➡️ 22:48 — The Heppner case (2026)

➡️ 26:26 — The Kovel doctrine (1961)

➡️ 28:14 — Incognito mode & deleted chats

➡️ 30:59 — The policy question

➡️ 34:00 — This is not a new problem

➡️ 37:05 — Are lawyers coal or horses? Jevons Paradox

Sponsor: Day One Law, a boutique corporate law firm that provides strategic legal counsel to startups, crypto projects, and Web3 innovators. ⁠You can get in contact with them via this link⁠: ⁠⁠https://www.dayonelaw.xyz/#contact.

Also: I'm re-launching the Law of Code newsletter as the world's shortest legal newsletter! You can ⁠⁠stay updated on emerging tech law for free here⁠⁠. https://www.lawofcode.fm/

Any feedback on this episode? Or how to improve the podcast? ⁠Click here⁠. https://forms.gle/W4d2a5aHuLJjuNdn7

Disclaimer: This podcast is for informational and educational purposes only and does not constitute legal or investment advice. Views expressed by guests are their own and do not necessarily reflect those of their employers. Listening to this podcast does not create an attorney-client relationship.