Jack Conte, CEO of Patreon, used his SXSW stage time this week to dismantle what he calls a “bogus” argument at the heart of the AI industry’s content strategy. As TechCrunch AI reports, Conte challenged AI companies’ reliance on fair use to justify training models on creators’ work without compensation, and he brought receipts.
His core point is sharp and hard to argue with: if fair use truly covers training data, why are AI companies cutting multimillion-dollar licensing deals with Disney, Condé Nast, Vox, and Warner Music? “If it’s legal to just use it, why pay?” Conte asked. The implication is clear: these deals exist because the legal ground is shaky, and the companies know it.
🔍 Why This Matters Right Now
The fair use debate isn’t academic anymore. It’s playing out in courtrooms (the New York Times v. OpenAI lawsuit), in Congress (multiple proposed AI training transparency bills), and in boardrooms where licensing deals are being negotiated behind closed doors.
What Conte highlights is a two-tier system forming in real time:
- Big publishers and studios get paid because they have legal teams and leverage
- Individual creators (illustrators, musicians, writers) get nothing because they don’t
This isn’t a new pattern. It mirrors how streaming platforms initially compensated major labels while independent artists earned fractions of a cent. But the scale here is different. AI training datasets consume the entire internet’s creative output, and the resulting models generate hundreds of billions in market value.
🎯 What Conte Actually Wants
Reading between the lines, Conte is positioning Patreon as a collective bargaining force. With hundreds of thousands of creators on the platform, Patreon could negotiate licensing deals with AI companies the way record labels negotiate with Spotify. That’s a smart business play disguised as advocacy, and it’s not a bad thing.
The creator economy needs aggregators who can represent individual voices at scale. No solo illustrator is going to negotiate a training data license with OpenAI. But a platform representing 250,000+ creators? That’s a different conversation.
📊 The Broader Industry Shift
Conte’s stance reflects a growing trend. Several developments are converging:
- Licensing is becoming the norm. OpenAI, Google, and Apple have all signed content deals in the past year. The question isn’t whether AI companies will pay: it’s who gets paid and how much.
- Regulation is catching up. The EU AI Act already requires training data transparency. US legislation is in draft stages.
- Creator tools are emerging. Services like Spawning and Kudurru let artists opt out of training datasets. But opt-out isn’t the same as getting paid.
What stands out here is Conte’s framing. He’s not anti-AI. He explicitly said the technology “is good, or it will be soon.” His argument is economic, not moral: if AI is the future, the people whose work built its foundation deserve a cut.
💡 Practical Takeaways
For AI practitioners and businesses, three things to watch:
- Budget for content licensing. Companies building on scraped data face increasing legal and reputational risk. Proactive licensing is cheaper than litigation.
- Expect platform-level negotiations. Patreon won’t be the last. Expect Substack, YouTube, and other creator platforms to push for collective deals.
- Fair use won’t hold forever. Courts haven’t ruled definitively yet, but the political and public sentiment momentum favors creators. Build your data strategy assuming you’ll need permission eventually.
Conte closed his talk with a line worth remembering: “Great artists don’t play back what already exists. They stand on the shoulders of giants. They push culture forward.” It’s a pointed reminder that the creative work AI companies consumed to build their models wasn’t just data: it was someone’s art, career, and livelihood.
The full details of Conte’s SXSW talk are available in the original TechCrunch AI report.