Tech Legal Brief #9 - Is AI Abolishing Copyright Protection?

New updates and thoughts on AI vs. copyright laws, why the use of AI should be banned in law school, and more.

Introduction

Today we will catch up on relevant developments in the Tech Legal space with the 9^th edition of Tech Legal Brief.

This concept is my rebellion against AI-generated news summaries. Instead of spending five minutes every other morning obsessively clicking links, reading headlines, and skimming clickbaity social media post and SEO-optimized articles, paid subscribers to Futuristic Lawyer dedicate 15 minutes every month or every other month to go deep on relevant new development in the areas I cover based on high quality sources and my personal style of writing. That is my business proposal.

These Briefs are not co-written by AI and they go much beyond surface-level news and headlines. Instead of observing the ripples on the sea, I dive straight in and search for pearls and treasures at the bottom of the ocean. Reading about tech legal news should not be a stressful experience or a boring drag, but a fun adventure.

On this note, here are the topics we will cover today:

- Is AI Abolishing Copyright Protection?

- The EU’s Emerging Approach to AI and Copyright Law

- Why the Use of AI Should Be Banned in Law Schools

- Tech Legal News (links)

You can find all prior editions of Tech Legal Brief here.

---

Is AI Abolishing Copyright Protection?

Some tech billionaire leaders have called for the abolition of intellectual property rights.

That is a radical proposal, obviously. Whether we like the idea of intellectual property rights or not, the society we live in would not be possible without them. For example, there wouldn’t be any incentive to invest millions upon millions of dollars in research for life-saving medicines if competitors could just reverse-engineer the end results and sell it for a cheaper price. Highly valuable trademarks such as Apple’s iconic logo or the distinct patterns on luxury goods from Gucci or Louis Vuitton would not retain brand authority and economic value if everyone were allowed to slap these trademarks on cheap products. Authors, musicians, and filmmakers would not earn revenue if their first buyer could resell the work as their own.

Still, the wish to abolish IP is not just limited to the eccentric minds of a few tech billionaires. In my view, the American BigTech industry is indirectly lobbying for it by making the argument that copyright laws do not apply to AI training. So far, the lobbying efforts appear to be working. Considering the American tradition of affinity and respect for IP laws, that is quite surprising. Only 12 years ago, a young open-source advocate by the name of Aaron Swartz tragically hung himself after facing federal charges of 35 years in prison for downloading and storing 4.8 million research articles from JSTOR. Now companies such as Meta, OpenAI, Google, Anthropic and x.AI are essentially claiming ownership to the internet’s entire body of knowledge, our cultural heritage, which includes public as well as proprietary data. Chances are they get away with it in America too.

The Empty Suits Win in America

We can’t meaningfully tackle the issue of copyright vs AI training without considering Trump’s tight-knit relationships with the group of famous tech leaders who symbolically lined up behind him on inauguration day. Each of them offered millions of dollars to Trump's inauguration fund, except for Elon Musk who spent hundreds and millions of dollars to ensure Trump was elected. If any of these lovely tech CEOs tried to curry favor with a President who is known for being extremely receptive to flattery and gifts, and kindly asked for an exception to copyright law for AI training, do you think such a change could be pushed through in the legislative machinery? I will go out on a limb and say it's probably not impossible.

If this situation makes you slightly angry, then great, mission accomplished. We need people who care and are slightly angry about this issue – without being so uncontrollably angry that they lose their minds – so they can practice peaceful resistance against a careless BigTech sector and a US political administration that seeks God’s blessing through cardinal sins.

Right now, BigTech policy looks to defeat US copyright law and tradition. In two recent court decisions – Bartz v. Anthropic and Kadrey v. Meta – the judges ruled that AI training is "transformative" and can be protected under the fair use doctrine in US copyright law. However, each decision had a very important limitation.

In Bartz v Anthropic, Judge Alsup ruled that Anthropic’s acquisition of pirated books from shadow libraries for the purpose of AI training was a copyright infringement. This could lead to a very substantial fine to Anthropic which will be determined in a separate trial. According to Judge Alsup, the company is storing more than 7 million pirated books in a “central library” and statutory damages can reach up to $150,000 per work.

Using pirated works in massive quantities to train AI models is the industry norm. In Kadrey v. Meta, Judge Chhabria ruled that Meta’s illicit downloading could be justified under the fair use doctrine. Meta won this case. However, Judge Chhabria underscored that the plaintiff did not present meaningful evidence on market dilution, and made it very clear that if the plaintiff had documented market dilution - e.g. declining sales due to cheap AI books - it would likely have changed the outcome:

“And, as should now be clear, this ruling does not stand for the proposition that Meta’s use of copyrighted materials to train its language models is lawful. It stands only for the proposition that these plaintiffs made the wrong arguments and failed to develop a record in support of the right one.”

For a more in-depth comparison of the two diverging court decisions see this post on Copyright Lately by Aaron Moss.

We are still very far from having any legal clarity on the state of training AI models through the lens of copyright law. To the contrary, we are in an early stage. Hopefully, some of the blockbuster cases which I have covered here in the past including The New York Times Company v. Microsoft Corporation, Sony Music Entertainment et al. v. Suno, Inc., and Authors Guild v. OpenAI Inc. can help to establish a precedent.

New cases are also continuing to roll in frequently. Joseph Savari law firm which represented the plaintiffs in Kadrey v. Meta has filed a new class action complaint against OpenAI less than a week after Judge Chhabria’s decision. Then we have Disney’s recent high-profile lawsuit against Midjourney. Copyright experts such as Andres Guadamuz (whom I previously interviewed on my podcast) appear to be very confident in the merits of this case.

The EU’s Emerging Approach to AI and Copyright Law

No matter what the outcome of the 44 copyright lawsuits against major lawsuits (and counting) will be in the US, and whether or not there will be a political intervention, I expect the EU courts to adhere strictly to the wording of existing copyright law. I also think that legal and political adjustments in the EU will favor artists and creatives over AI companies.