USC law professor Jonathan Barnett says a secure baseline of intellectual property protection, with appropriate adjustments, would reflect value contributed by both AI technology and content creators.
The rapid adoption of artificial intelligence services has reawakened the tension between tech and content over copyright in the digital economy. Protecting intellectual property rights while allowing tailored exceptions will be critical to treating both sides fairly.
Tens of lawsuits over AI model developers’ use of copyright-protected content are proceeding through federal courts. Content owners have received favorable rulings in some cases.
The US Copyright Office released a report last month in which it took the middle-of-the-road position that not all uses of copyright-protected content by AI models qualify for the fair use exemption. The next day, office chief Shira Perlmutter was fired.
Today’s debates over copyright and AI recall the debates over copyright and piracy that followed file-sharing platform Napster’s launch in 1999. Then and now, the conflicting positions of content and tech reflect divergent business models.
Content seeks secure IP rights, which support revenue streams that sustain investments in new releases. Tech favors insecure IP rights, which reduces the costs of acquiring creative fuel for the devices and apps it sells to users.
Much of the academic and advocacy communities endorse positions that align with the views of several large tech platforms. This accidental alliance has cultivated an intellectual climate that normalizes copyright infringement as part of a seemingly attractive “sharing economy” in which ownership is a defunct concept from the analog world.
Yet this utopian view reflects a one-way business logic that consistently favors digital aggregators over content originators. Reducing the cost of acquiring content maximizes a platform’s user base, which maximizes its profits on ads and other proprietary services.
That means content producers earn little for the value they contribute. Generative AI accelerates this effect, as most users never reach the original content reflected in a chatbot’s response.
Tech’s “IP-lite” vision has largely prevailed in legal disputes over copyright in the digital ecosystem. Following the judicial shutdown of Napster in 2001, some platforms doubled down and adopted a risky strategy of “take, then litigate.”
The gamble paid off. In cases involving Google Books, Google Images, and YouTube, platforms successfully defended mass copying, or practices that facilitate mass copying, through push-the-envelope applications of exemptions to infringement liability.
Fair use doctrine has been the key. Tech platforms argued, and courts mostly agreed, that the exemption should be broadened to facilitate the digitization of creative content. This once controversial application of fair use is now widely accepted.
This expansive application of fair use departs from historical practice. Until roughly the mid-2000s, courts had generally reserved the exemption for limited uses such as excerpts. Further, this generous understanding stands in tension with US Supreme Court precedent, which has held that fair use doesn’t apply when there is significant commercial harm to the copyright owner.
The Supreme Court reminded the judiciary in 2023 that fair use is an exception—not the rule—in upholding an individual photographer’s infringement claim against the Andy Warhol estate.
The fair use consensus that predominates in IP policy circles typically dismisses calls to narrow the exemption as a thinly disguised “cover” argument for media interests. But the economic logic behind fair use’s expansion shows that the equities often run in the opposite direction. In ad-based digital markets, lifting copyright restraints on unauthorized usage effectively transfers wealth from originators that produce content to large platforms that aggregate content to promote ads and other services.
The rise of the creator economy has challenged the fair use consensus by showing how ownership empowers individual artists to capture value in creative markets. This principle motivated the original copyright statute, the UK’s Statute of Anne enacted in 1710, which inspired the Intellectual Property Clause in the US Constitution. This may explain the backlash in the UK against recent proposals to compel copyright owners to opt out of a text-and-data-mining exemption for AI model developers.
Tech and like-minded commentators say narrowing exemptions to infringement liability will imperil the AI ecosystem. But this argument is belied by the evolution of digital content markets that have thrived by renewing property rights, not eliminating them.
Although tech secured legal wins in the post-Napster era that largely shield mass copying through online platforms, successful digital content markets nonetheless adopted technological equivalents that partially offset this copyright deficit.
Without those access controls, streaming platforms couldn’t secure the subscription revenue that keep them in business. The dismal fortunes of most “free” ad-supported news platforms show that property-free environments are profit-free environments that lack viability.
Unlike IP-free frameworks that arbitrarily favor aggregators over originators, a meaningful IP backstop reflects a two-way business logic that is fair for everyone. Respecting IP rights maintains the revenue flows that sustain investment in higher-value original content.
Just like markets in physical goods, markets in creative goods can’t function without meaningful property rights—whether enforced legally or technologically—to reflect the value of creative works, which generates revenue that support new content development.
This isn’t to say copyright should be enforced reflexively. Nuanced exemptions play a role in a balanced IP framework. AI developers raise legitimate concerns that sweeping copyright enforcement could hinder AI-enabled technologies that enable new creative possibilities.
Yet only a secure baseline of IP protection, with appropriate adjustments, can enable markets to craft contractual and technological mechanisms that reflect the value contributed by both tech and content to the AI ecosystem. Anything else would be unfair.
This article does not necessarily reflect the opinion of Bloomberg Industry Group, Inc., the publisher of Bloomberg Law, Bloomberg Tax, and Bloomberg Government, or its owners.
Author Information
Jonathan Barnett is an author, law professor, and director of the media and technology law program at the University of Southern California’s Gould School of Law.
Write for Us: Author Guidelines
To contact the editors responsible for this story:
Learn more about Bloomberg Law or Log In to keep reading:
Learn About Bloomberg Law
AI-powered legal analytics, workflow tools and premium legal & business news.
Already a subscriber?
Log in to keep reading or access research tools.