Has Copyright Gone Wrong?

In partnership with

In every course that I’ve taught at Georgian College, I dedicate at least one lecture to consumer privacy, regional digital content legislation, and global internet-related regulations. I get excited about it. The students, not so much.

I let the class know that it’s thanks to the European Union’s General Data Protection Regulation (GDPR for short) that we must all accept cookies (it’s okay, cookies, I accept you) on every fucking website we visit.

I explain how, in most jurisdictions, the only “real” protection for anything online is long-standing copyright laws; specifically, the application and understanding of what constitutes “fair use” of a piece of content or art.

And I will wax poetic (possibly even drone on) about how copyright has been challenged regularly since the birth of the World Wide Web at CERN.

Notably, there was Sean Parker’s and Shawn Fanning’s Napster, the seminal service that allowed MP3 file searching and file sharing across the internet, copyright be damned.

The music industry, and oddly Lars Ulrich, the drummer of Metallica, would not let illegal file sharing stand. The courts agreed at the time that file sharing was not fair use, and Napster was forced to shut down.

It took Apple and its deep pockets to license most of the available music on the planet and agree to pay labels for downloads to set a new precedent on how music could be distributed. Arguably a legitimate rerendering of what the Seans (or the Shawns?) were trying to do with Napster.

In Canada, the Professional Writers Association of Canada (PWAC) fought the Globe and Mail for over a decade before it won a quietly historic victory. However, they didn’t win much. Somehow, both sides claimed victory.

Big tech and media’s dance with copyright isn’t often harmonious.

Google has consistently challenged copyright norms and has been sued a lot because of it.

Like, a lot, a lot.

So much so that the internal training team at Google worked hard to teach employees to avoid including specific words in emails, disable message history for sensitive conversations in chat, and add lawyers to any email threads that might prove to be potentially legally contentious in the future.

(I remember the training.)

Google’s legal caution was not just for copyright concerns, but under the auspice that any email or chat might end up in a court case one day. For instance, you didn’t want to be the Googler saying in an email that you fixed the ad auction so that Google could hit its quarterly sales targets.

The history of Google’s copyright battles reads like a history of the internet and our changing perspectives as consumers of online services over the last couple of decades.

Agence France Presse (AFP) sued Google for copyright infringement in federal court in the District of Columbia, a case which Google settled for an undisclosed amount.

In 2005, the Authors Guild and the Association of American Publishers (AAP) sued Google twice, alleging copyright infringement for its Google Books project. Google’s scanning and indexing of the books was deemed fair use.

Viacom sued Google’s YouTube in 2007, claiming the company had engaged in “brazen” and “massive” copyright infringement by allowing users to upload and view hundreds of thousands of videos owned by Viacom without permission. The judge in that case granted summary judgment in favour of Google (YouTube) as well.

Mian Mian, a Chinese Post 70s Generation novelist, sued Google in 2009 after the company scanned her books for its online library. She demanded CN¥61,000 (just over $400 in today’s dollars) and a public apology. Google later removed the book from its library and paid her the requested tiny settlement. In this case, the artist won. But again, they didn’t win much.

You get the idea.

Google has an army of lawyers. And most of the time, those lawyers win more than they lose.

In cases where Google can’t win, it tries to settle.

That’s the advantage you have when you are a company that is worth over two trillion dollars: you can pay your way out of a lot of problems.

Today, Google is being challenged for its use of trademark material in its AI products.

Last year, Google was fined €250m by French regulators for breaching an agreement related to paying media companies for their content. They said that Google violated the intellectual property rights of news publishers. The regulator also cited concerns about Google’s AI service, saying that Google’s Gemini was trained on content from publishers without their consent.

And a group of visual artists sued Google, alleging that Google’s AI-powered image generator, Imagen, was trained on their copyrighted content without authorization.

Google is not alone.

Open AI, Microsoft, Meta, Midjourney, Perplexity, Cohere, Anthropic, and other AI companies are facing similar legal action. The argument these AI giants use kinda sounds similar: training AI isn’t a copyright violation, it’s an exercise in learning. The suggestion is that building an AI model is like using textbooks to teach an average college student. They argue that it amounts to a transformative application of the content that is covered by the trademark laws’ definition of fair use.

In a modern echo of Google’s book scanning case, a judge recently ruled that Anthropic’s use of millions of books to train its AI models qualifies as fair use. Anthropic might still have to pay for the fact that it didn’t buy or pay for a lot of those books, but that’s another story.

Meta won a similar legal case.

However, in that Meta case, the judge may have given authors and artists a golden ticket. Although he said the authors’ case was wrongly argued, he gave clues on how a better argument might be constructed.

The Donald stepped into the AI copyright debate a couple of weeks ago, making a statement that AI companies, “can’t be expected to have a successful AI program when every single article, book, or anything else that you’ve read or studied, you’re supposed to pay for,” Trump said. “We appreciate that, but just can’t do it— because it’s not doable.”

Like on most days, The Donald is wrong.

Tracking content use and paying artists for their work is something that the music industry has done for decades. This was how and why they were able to win against Napster in the first place.

Technology companies, with vast resources and enormous amounts of compute to generate AI models, are best situated to use those resources (possibly even leveraging AI) to identify where they are using copyright material, are they not?

The battle is far from over.

More artists and publishers are likely to challenge how their content got into AI models.

And there are still many things to debate regarding the output of AI.

Maybe training AI is fair use, but is generating work that is too similar to an artist’s original work fair?

If it looks like a duck and sounds like a duck, is that a violation of the duck’s copyright?

Quack. 🦆

Time and a lot of lawyer fees will help decide how it all flushes out.

In the interim, Sony is partying like it’s 1999 by suing Napster again.

This time, Napster is being sued for not paying its bills.

Maybe AI companies will also need to pay their bills, because no artist’s work is free.

Finally, a powerful CRM—made simple.

Attio is the AI-native CRM built to scale your company from seed stage to category leader. Powerful, flexible, and intuitive to use, Attio is the CRM for the next-generation of teams.

Sync your email and calendar, and Attio instantly builds your CRM—enriching every company, contact, and interaction with actionable insights in seconds.

With Attio, AI isn’t just a feature—it’s the foundation.

• Instantly find and route leads with research agents

• Get real-time AI insights during customer conversations

• Build AI automations for your most complex workflows

• Join fast growing teams like Flatfile, Replicate, Modal, and more.

Start for free today.