Are AI Copyright Lawsuits Ending?
Instead of battling it out in court, some of the biggest fighters against GenAI training are suddenly all bending the knee towards licensing, including the music industry, New York Times, and Reddit.
We are starting to watch the beginning of the end… for AI copyright lawsuits.
All the screaming about data theft, copyright infringement, scraping, suddenly, it’s going quiet, because companies aren’t trying to fight anymore.
They’re all starting to… license.
In the last seven days, THREE major shifts happened with music labels, The New York Times, and Reddit, all pointing to one thing:
Companies are moving from “stop using our data”... to “okay its fine, but pay us, and give us a seat at your table.”
Video Debrief Here:
Even though the law around using people’s data for AI training still isn’t fully settled, moves like these do set legal precedent. And if every major corporation starts playing nice all at once, it’s hard not to see that as a turning point. These are the same companies that spent years filing lawsuits… and now, almost overnight, they’re changing course.
FIRST EXAMPLE: The Music Industry
One of the loudest voices in the fight against generative AI.
Universal Music Group, Warner Music Group, and Sony Music Entertainment have been in a legal standoff with AI audio startups like Suno and Udio for over a year now.
Their claim is that these tools illegally trained on copyrighted songs and were generating music close to the originals, sometimes even mimicking the voices of real artists.
Last year, the labels took them to court, filing high-profile copyright lawsuits. It felt like the start of a long legal war over how (or whether) AI models could be trained on copyrighted material like music.
But this week, Bloomberg dropped a report that changed the tone entirely.
The same record labels are now in talks with Suno and Udio. Not to shut them down, but to license music to them and negotiating equity stakes in the startups themselves.
What a 180?.. So what’s going on here?
I used to work for one of these record labels, Sony Music, in music copyright compliance. Let’s talk about why music rights are especially compliance and how I’m a bit confused these labels can even really have these conversations.
In the U.S., every piece of music has two main copyrights:
The specific sound recording of the song (often owned by the label, that ones easier)
The musical composition (the words, the melody, typically owned by songwriters or publishers, like a lot of them. It’s very common for 10+ writers to have a slice of this)
So even if all these major labels to license the recording and their portions of that second copyright, the underlying song might have multiple other rights holders who still need to sign off.
So negotiating a blanket license isn’t as simple as them just making a decision for everyone, because its likely every song has dozens of multiple copyright holders that need to sign off.
If these licensing deals go through, it would most likely set a precedent for how music rights would be handled in future AI training for musicians if they agree with it or not.
EXAMPLE 2: The New York Times Signs Licensing Deal With Amazon
(while actively suing OpenAI)
Last year the New York Times made headlines by filing a major lawsuit against OpenAI and Microsoft, accusing them of using Times articles to train ChatGPT without permission. It was the first big case of a legacy media outlet pushing back on AI training data, and they went all in — not just asking for money, but demanding OpenAI stop using their content completely and delete anything already trained on it. That case is still ongoing.
But this week, The Times signed a licensing deal, not with OpenAI, but with Amazon. The agreement gives Amazon access to a wide range of Times content, including news articles, NYT Cooking recipes, and sports reporting.
So here’s the contradiction: on one side, The Times is suing one company for using its content without permission, and on the other, it’s actively selling that same content to another. It’s not about whether AI can use the data... it’s about who gets to profit from it.
And it’s also positioning the the Times as a gatekeeper. Because they’re saying:
If you want to build your model on our journalism,
You pay. You get permission.
And we decide how our content shows up inside your product.
Third Example: Reddit
The third example is Reddit filed a lawsuit against Anthropic this week, and it tells a slightly different story. Unlike the music labels or The New York Times, Reddit never claimed copyright infringement and saying they stole their content. Instead, they say Anthropic violated licensing terms (meaning, the issue isn’t that the data was used, it’s that it was used without paying for it).
And that subtle distinction says a lot. Reddit already has licensing deals in place with companies like OpenAI and Google. So the message here isn’t “don’t train on our data,” it’s “you can, but not for free.”
Why is This All Happening?
Now this is alot to digest but I want to point to one possible catalyst for all of these legal shifts.
A few weeks ago U.S. Copyright Office released a pre-publication version of its long-awaited decision on whether training AI on copyrighted data is legal. The document is over 100 pages and makes it clear that, in most cases, using copyrighted material to train AI models would likely be considered infringement.
The timing of it was super weird though, because the report published as a “pre-publication,” which isn’t standard practice, and the Register of Copyrights who oversaw it was dismissed by the Trump administration the very next day. So it’s unclear whether this guidance will hold or be officially adopted.
But I do wonder if this shift we’re seeing from lawsuits to licensing is a strategic response to that report to urgently play nice. Companies may be reading the room and realizing that even if the legal precedent isn’t fully locked in, the direction is clear.
Fighting AI companies in court hasn’t delivered fast results, but licensing gives them control, payment, and a seat at the table.
And if that becomes the norm, it may not matter what the law says… because the market will have already decided.
PS: Tomorrow I will publish my normal “all the updates for the week”. If you want to see all of them now, watch the full youtube video. :-)
PSPS: If you want to support to help fuel me to keep the videos coming, please considering treating me to coffee (or matcha!!) https://coff.ee/veronicahylak