This episode is a discussion of host Peter Csathy's new article titled, "Court Rejects AI's Fair Use: Now What?" (generated using Google NotebookLM). Csathy's article discusses all the issues raised by (and ramifications of) Judge Stephen Bibas's recent game-changing rejection of "fair use" as a defense to unlicensed AI scraping in the Thomson Reuters v. Ross Intelligence case: (1) now what will generative AI companies do (how will they act)?, (2) how would courts even begin to assess damages in this "horse is already out of the barn" scenario where the entire internet has already been scraped, (3) how would the Supreme Court rule when that AI training "fair use" issue inevitably reaches them?, and (4) would Congress itself create a new "fair use" exception to permit unlimited unlicensed AI training on copyrighted works (especially in the face of the new co-sheriffs in town, President Trump and Elon Musk)?

Csathy approves the contents of this episode, and believes it is an accurate reflection of the issues (and his overall analysis). BUT there is one part of the discussion near the end of the episode that focuses on training on "synthetic data" - and Csathy strongly disagrees on that point.

Sign up for the companion "the brAIn" newsletter via this link.
Check out Peter and his firm Creative Media
Check out Peter's LinkedIn bio here.
And send feedback to bizdev@creativemedia.biz.