When the popular file-sharing platform Napster received a court order to shut down in 2000, lawyers representing the recording industry successfully made the argument that the startup infringed on their copyrighted works.
Now, as a rush of entrepreneurs and large companies look to cash in on the rise of generative artificial intelligence, concerns that the technology industry is playing fast and loose with intellectual property laws are starting to resurface.
A growing number of music labels, comedians, computer programmers, and others are worried that AI models are being trained on their intellectual property, without proper recognition or compensation.
This was the focus of a panel discussion hosted by Techstars this week in Seattle titled “Navigating the IP Landscape in AI.”
Panelists included Jordan Ritter, co-founder of Napster; Heather Johnson, director of label operations at FiXT Record and Publishing; and Gary Greenstein, partner at Wilson Sonsini. Techstars managing director Marius Ciocirlan moderated.
In the case against Napster, the startup’s founders and supporters argued that the recording industry was misusing copyright law to stifle technology, restricting the internet’s role as a free platform for sharing music, videos, and other types of content.
“If we all had properly respected copyright law, we might not have digital music today,” said Ritter, speaking during the panel event at Startup Hall at the University of Washington. “We might be still stuck with CDs.”
The discussion touched on the parallels between the Napster era and now; legal approach AI entrepreneurs can take; and the ownership structure of AI-generated content. Read on for highlights from the panel, edited for brevity and clarity.
Digital streaming vs. AI content era
Ritter: “What made Napster a different situation was that there was an economic market force involved. We were all being forced to pay $20 to get the one song we wanted out of the 12 that we didn’t really care about. People were fed up. And so we wanted something different. The industry wouldn’t give it to us. We ended up building it ourselves. That was why that disruption came to be.
I think this one is a voluntary disruption, different in that way. It’s no less impactful, maybe more impactful, but it’s not driven by an economic factor.”
Greenstein: “One thing you will see similar to Napster is an effort to preserve lawful monopoly. The control that [the recording industry] wants to exercise is legitimate, a right granted to them. And so all of you who are creating companies, I strongly encourage you — in addition to hiring a good lawyer — to think about how you’re operating your business.”
Forgiveness or permission?
Greenstein: “For generative AI, the manner in which you’re training your algorithms and going out and getting data, it’s going to be better if you license that content. You can go to someone who has lawfully acquired the rights, who has already built the underlying database, or a copyright owner, record company, film studio, book publisher, and get authorization.
“If you eschew all of that and decide, ‘you know what, I don’t have the time and money, I’m going to move fast break things. If I get big enough to get sued, god bless me, I’ll go and raise money and I’ll settle with the copyright owners.’ Those are approaches that people take. Forgiveness or permission?”
Johnson: “This is one of the instances where the music industry could probably learn from what happened with Napster, and maybe be much more open to approaching new AI technologies and saying, ‘Hey, can you talk about licensing with us and work out some sort of fair market price?’ There’s nothing that really defines that right now. But having those types of conversations and negotiations as a starting point, as opposed to relying on lawsuits or statutory rates that come from laws that get passed.”
Artist perspective
Johnson: “We have some artists who are fascinated with the new AI technologies and see so many opportunities. We’ve had to define rules on what we’re willing to accept. Currently, with the Copyright Office, you can’t register something not human-authorized. So we can’t actually protect what they’ve created if there’s too much — or a lack of — human authorship in it. Some of them are terrified by it, worried that they’re going to be replaced, or that someone is going to copy their voice.”
An entrepreneur’s perspective
Ritter: “The gleeful abandon we had back in the day around Napster and lack of consideration towards copyright is not something I share these days as much. (Ritter is also the co-founder of personal AI assistant startup Augment). But I do think that restriction and legal stuff is the enemy of innovation and ingenuity.”
Should AI creations be copyright protected?
Greenstein: “If it’s created by AI, it should not be protected by copyright. Intellectual property laws are designed to promote science and the useful arts. There’s an economic incentive inherent in that if a person is exercising their brain coming up with inventions or creating original works of authorship, giving them protection for limited periods of time allows them to exploit that and generate revenue. The incentive in our system is they will continue to create. AI doesn’t need an incentive. The people who invest in the computer may need an incentive, but the output that is the product of generative AI should not be subject to copyright protection.”
