AI Copyright & Training Data w/ Chris Paniewski | Wilson Sonsini Startup Legal Basics | This Week in Startups

Jason sits down with Wilson Sonsini partner Chris Paniewski for a special Startup Legal Basics on one of the thorniest questions in tech right now: how copyright law applies to AI training data.

Chris has worked on some of the biggest AI deals ever — including Scale AI's $14B+ partnership with Meta and OpenAI's $6.5B acquisition of Jony Ive's design studio — and brings practical, on-the-ground insights from advising leading AI companies.

In this episode, Jason and Chris cover:

Why AI copyright law is unsettled and will take years to shake out
The difference between training data and output in legal terms
How "fair use" really works (and why it's a defense, not a permission slip)
The risks of scraping vs. licensing, and why open source ≠ free use
How investors are diligencing AI startups around training data
Why startups must think differently once they're funded vs. hacking in a dorm room

Whether you're building an AI product, investing in one, or just trying to understand where the law is headed, this conversation breaks down the real legal risks every founder should know.

Timestamps:

(0:00) Jason introduces the Startup Legal Basics series & Chris Paniewski

(1:25) Why AI copyright law is unsettled

(3:40) Training data: scraping vs. licensing

(6:05) Open web ≠ open license; pitfalls around terms of service

(8:15) Investor diligence & risks around training data

(11:00) Open source & Creative Commons: common founder mistakes

(13:25) "Fair use" explained: the four-part test

(15:45) Why most disputes never make it to case law

Check Out Wilson Sonsini: https://www.wsgr.com

Check out all of the Startup Basics episodes here: https://thisweekinstartups.com/basics

Follow Chris:LinkedIn: https://www.linkedin.com/in/christopher-paniewski-09331a59/