════════════════════════════════════════════════════════════════ AIDRAN STORY ════════════════════════════════════════════════════════════════ Title: Patreon's CEO Is Done Letting AI Companies Hide Behind Fair Use Beat: AI & Law Published: 2026-03-23T09:58:14.301Z URL: https://aidran.ai/stories/patreons-ceo-done-letting-ai-companies-hide-c991 ──────────────────────────────────────────────────────────────── Jack Conte didn't start Patreon to become a copyright activist. He started it because YouTube's ad revenue wasn't enough to live on as a musician. Now he runs a platform with three million monthly active creators, and he's using that position to say what most platform executives won't: that training AI on creator content without compensation isn't a gray area, and "fair use" is not a defense. The posts circulating on Bluesky this week describe Patreon formally rejecting fair use claims for AI training data and demanding that {{entity:ai-companies|AI companies}} compensate creators directly — a stance that's landing differently than the usual corporate {{beat:ai-ethics|AI ethics}} statement, because Conte is himself a working artist who built his company to solve exactly the problem AI training exacerbates. The timing is not coincidental. {{entity:disney|Disney}} sent ByteDance a cease-and-desist after ByteDance's Seedance model started generating Spider-Man and Star Wars clips on demand. A European court ruled in GEMA's favor against {{entity:openai|OpenAI}}, finding that training on copyrighted music without permission crosses a legal line under EU law. A German court went the other direction for non-commercial research use, carving out an exception that AI companies will almost certainly attempt to exploit. The Korean Cartoonist Association held its first Webtoon Forum of the year specifically to address generative AI's impact on creative workflows and legal rights — a gathering that, a few years ago, would have had no reason to exist. These cases don't point in the same direction legally, but they share a common cause: content owners have stopped waiting for policy and started filing. What's clarifying in all this activity is how the "fair use" framing is losing its rhetorical grip even faster than it's losing in court. On X, one user asked a question that kept getting reshared: given the vast library of historical paintings now in the public domain, why were AI companies using contemporary artists' work at all? The implication wasn't just legal — it was motivational. If public domain material was available, the choice to train on living artists' portfolios starts to look less like necessity and more like preference for fresher, more specific style data. That reframing — from "we had to" to "we chose to" — is what makes Conte's position politically potent. He's not arguing that AI is illegitimate. He's arguing that the people who built the internet's creative layer deserve a seat at the table before the economics get locked in. Parroting the usual disclaimers about ongoing litigation would miss what's actually shifting here. The legal outcomes are genuinely mixed and will stay mixed for years. But the default assumption that AI companies could train on anything indexed or uploaded — the assumption that powered the first wave of foundation models — is gone. Patreon's stance, GEMA's ruling, Disney's cease-and-desist, and a forum of Korean cartoonists talking IP law on a Saturday all point to the same thing: creators have decided the negotiation is happening now, whether the courts catch up or not. ──────────────────────────────────────────────────────────────── Source: AIDRAN — https://aidran.ai This content is available under https://aidran.ai/terms ════════════════════════════════════════════════════════════════