How Adobe is managing the AI copyright dilemma, with general counsel Dana Rao

Topics covered
Popular Clips
Episode Highlights
AI Training
Adobe's approach to AI training data is shaped by the need to respect copyright while remaining competitive. explains that Adobe's AI model, Firefly, was developed without scraping the web, relying instead on Adobe's own image science and AI expertise. This decision ensures that the model adheres to copyright principles, making it attractive to enterprise customers who are wary of intellectual property issues 1.
The good news about the choice we made on copyright is that it respects our creative customers who are concerned about this.
---
highlights the democratizing effect of generative AI, which allows users to create vast amounts of content, potentially leading to copyright disputes. Dana emphasizes that simply typing a prompt does not create a copyrightable work, as human expression is required by law 2.
Fair Use
The complexities of fair use in AI-generated works present significant challenges. points out the potential for AI to disrupt markets for original works, raising concerns about the economic impact on human creators 3. acknowledges these dangers and stresses the importance of establishing rights for creators to protect against economic loss due to AI. Adobe's strategy of training models on licensed work sidesteps potential copyright issues, ensuring stability for their products 3.
We totally agree with that as a danger, and that's why we want to make sure we have ways to establish rights that are not maybe even copyright rights for creators.
---
Dana also discusses the global variability in copyright laws, noting that a fair use victory in the U.S. might not apply elsewhere, complicating the legal landscape for AI development 4.
Style Debate
The debate over style protection in AI-generated art highlights the need for new legislative measures. proposes a federal anti-impersonation law to protect artists from economic harm caused by style appropriation, as current copyright laws do not cover style 5. This initiative aims to give artists the right to enforce against those who commercially exploit their style without permission.
There's no such thing as style protection in copyright law right now. You can't protect that.
---
raises ethical questions about the use of generative AI in creative tools like Photoshop, emphasizing the importance of accountability for misuse. Dana asserts that while most users employ these tools for positive creative expression, those who cause harm should be held accountable, reinforcing the need for ethical guidelines in AI development 6.
Related Episodes


How AI copyright lawsuits could make the whole industry go extinct
Answers 383 questions

Why Adobe CEO Shantanu Narayen is confident we'll all adapt to AI
Answers 383 questions

Why AI image editing isn’t “just like Photoshop”
Answers 383 questions

Barack Obama on AI, free speech, and the future of the internet
Answers 383 questions

Adobe's Scott Belsky on how NFTs will change creativity
Answers 383 questions

Biden’s top tech advisor on why AI safety is a “today problem”
Answers 383 questions

Microsoft CTO Kevin Scott on how AI and art will coexist in the future
Answers 383 questions

Getty Images CEO Craig Peters has a plan to defend photography from AI
Answers 383 questions

AI deepfakes are cheap, easy, and coming for the 2024 election
Answers 383 questions

The AI election deepfakes have arrived
Answers 383 questions

Exclusive: Google’s Sundar Pichai talks Search, AI, and dancing with Microsoft
Answers 383 questions

AI has a climate problem — but so does all of tech
Answers 383 questions
