Version 4

What

Why it matters

Open questions

• Will the opt-out authors pursuing a jury trial [7] recover more than the ~$3,000-per-work class figure, or will courts apply the same training-as-fair-use reasoning from Bartz to cap their claims?

• Does Google's DMCA theory — that anti-scraping technology constitutes copy protection — survive appellate review [14][15]?

• How will courts define the line between capturing an author's 'broad qualities' and impermissible style mimicry, given that ChatGPT's new policy draws that line without judicial endorsement [16]?

• Does Saflor v. Memes Apps LLC establish that commercial AI exploitation of viral content requires a license, or will courts treat widespread prior viral sharing as severing the copyright holder's commercial claims [17][19]?

Narrative

Timeline

• 2017: Elmer Saflor's 'Running Away Balloon' comic goes viral, generating widespread unpaid copying online. [17]
• 2025-12: Google files a DMCA lawsuit against SerpApi, characterizing its commercial scraping of Google search results as 'parasitic' circumvention of copy protection. [10][11]
• 2026-01-12: Authors Alliance reports updated opt-out and objection deadlines in Bartz v. Anthropic, and notes a new judge has been assigned to the case. [4]
• 2026-01-20: Society of Authors (UK) publishes Anthropic's list of works used in training, providing member information on the settlement. [5]
• 2026-02-21: SerpApi files a motion to dismiss Google's lawsuit, arguing Google does not own publicly accessible search results. [12][13]
• 2026-05-14: Authors who opted out of the Anthropic class settlement file a separate copyright lawsuit requesting a jury trial. [7]
• 2026-07-17: Writer Beware covers the opt-out authors' separate litigation under the heading 'The Opt-Outs Strike Back.' [8]
• 2026-07-20: A federal judge grants final approval to Anthropic's $1.5 billion settlement with authors, the largest copyright settlement in U.S. history. [20][1]
• 2026-07-21: Reporting confirms only 350 authors opted out, that the court ruled AI training was fair use but piracy likely was not, and that individual payouts were approximately $3,000 per work. [2][21][22]
• 2026-07-27: A federal court rejects Google's key DMCA claims against SerpApi, ruling that anti-scraping technology does not constitute copy protection; Google announces it will appeal. [14][15][23][24]
• 2026-07-27: ChatGPT begins refusing direct requests to mimic named authors' styles, extending a prior restriction on living authors to deceased ones. [16]
• 2026-07-27: Artist Elmer Saflor sues Memes Apps LLC, alleging its paid AI meme-generation service commercially exploits his copyrighted 'Running Away Balloon' comic without authorization. [17][18][19]
• 2026-07-27: NPR reports authors have mixed feelings about the Anthropic settlement, with some welcoming the piracy finding and others viewing the per-work payout as inadequate. [3]

Perspectives

Tensions

• Authors opposing the settlement argue the ~$3,000-per-work payout and high attorney fees make the deal inadequate; the court and the majority of class members accepted the terms as sufficient. [2][3]
• The court treats AI training as fair use but direct piracy as likely not; opt-out authors pursuing a jury trial argue the fair use ruling set the compensation bar too low and are testing whether they can recover more. [2][7]
• Google argues its anti-scraping technology constitutes DMCA-protected copy protection; SerpApi argues Google does not own publicly accessible search results, and the trial court sided with SerpApi. [10][14][15][12]
• ChatGPT's new policy draws a line between capturing 'broad qualities' and direct imitation, but authors may argue the distinction offers no meaningful protection for their distinctive voice. [16]
• Saflor argues commercial AI exploitation of viral content requires authorization; Memes Apps LLC has not yet responded on the public record, leaving unresolved whether widespread prior viral sharing severs the copyright holder's commercial claims. [17][18][19]