The media world is splitting in two over how to handle generative AI scraping its copyrighted work. News organizations are heading down one of two paths: they’re either suing the AI companies in expensive, high-stakes lawsuits or they’re cutting licensing deals. This is the core fight over AI content rights right now, and it’s going to define digital publishing and IP for a long time. The big question is, can publishers actually get control back, or are they just walking into a legal and financial quagmire?
Key Takeaways
- Big names like The New York Times are suing AI developers for copyright theft, and it’s a trend that’s picking up speed in media law.
- The alternative is to make a deal, like The Associated Press did with OpenAI, which gives publishers a new revenue source and some say in how their work trains AI.
- Nothing’s certain legally, and the big court cases that will define ‘fair use’ and data scraping probably won’t set clear precedents until 2026 or beyond.
- Every publisher now has to weigh the immediate costs of a lawsuit against the long-term risks of a licensing deal to figure out how to protect their intellectual property.
- AI companies are now looking for partners to make their data sources legitimate, which could mean a move toward more collaboration instead of just scraping.
| Factor | Litigation (e.g., NYT) | Licensing (e.g., AP) |
|---|---|---|
| Primary Goal | Make AI developers pay for infringement | Get paid and have some control over use |
| Key Action | Suing for copyright infringement | Signing deals for content access |
| Timeline | Long, multi-year cases, clear rules maybe by 2026 | Immediate revenue from deals made in 2023 |
| Financial Aspect | Huge legal costs, hoping for big damages | A new way to make money |
| Control/Influence | Trying to force a change in how AI gets data | Gets you attribution, some accuracy, and a cut of revenue |
| Risk/Uncertainty | Very expensive and you might still lose | Makes AI scraping seem OK, might devalue your own work |
Litigation Mounts Against AI Developers
Talk is over. A growing number of news outlets are now suing. The New York Times really got the ball rolling in December 2023 with its massive lawsuit against OpenAI and Microsoft, arguing their AI models were illegally trained on millions of its copyrighted articles. Filed in the Southern District of New York, the suit is a direct attempt to make AI developers pay for what the Times calls systematic theft. The paper’s point is that these AI models aren’t just learning, they’re spitting out its journalism almost word-for-word, creating a direct competitor that damages its own subscription business. This kind of aggressive legal move shows just how fed up publishers are with watching their work get used to build someone else’s product for free. And they aren’t alone. Reuters reported in early 2026 that a bunch of smaller news agencies are getting their own lawsuits ready to claim damages for having their archives scraped without permission. At the heart of it all is a simple question that’s now being tested like never before: does training an AI on copyrighted material fall under “fair use”? Attorneys who practice media law are saying not to hold your breath, as these cases will likely take years and could easily end up in front of the Supreme Court.
The Rise of Licensing Agreements
Instead of suing, some publishers are just making deals. The Associated Press (AP) jumped on this early, announcing a licensing agreement with OpenAI way back in July 2023. This arrangement gives OpenAI access to the AP’s gigantic news archive to train its models. In return, the AP gets a new revenue stream and, just as important, a seat at the table to influence how its journalism is actually used. It’s a pragmatic move, get paid now instead of fighting a long, expensive war you might lose anyway. German media giant Axel Springer followed suit, signing a global partnership with OpenAI in December 2023 that allows the AI to summarize articles from its brands, including Politico and Business Insider, directly in its products. These deals typically come with requirements for giving credit and sharing revenue, creating a blueprint for how tech and media can work together. Publishers making these agreements want the money, sure, but they’re also trying to make sure the AI represents their reporting accurately and with the right context. I think this kind of proactive deal-making is the smarter play for many. It’s certainly better than pouring money into a legal fight with no guaranteed winner.
Implications for the News Industry
So, these two paths, suing or signing, create a messy future for the news business. If the lawsuits pay off, AI companies might have to completely rethink how they get their data and could be forced into compensation models that actually pay creators fairly. But if licensing becomes the norm, news archives suddenly become a hot commodity for AI training, which could be a lifeline for publishers that are already struggling to make ends meet. There’s a catch either way. Litigation is a money pit with absolutely no guarantee of a win. At the same time, licensing deals might bring in cash today but could end up legitimizing the very AI tools that devalue original reporting tomorrow (a classic devil’s bargain). The industry’s choices here are fraught, balancing the need to make payroll this quarter against protecting the long-term value of their journalism and IP. What happens in these fights will set the terms of engagement between tech companies and content creators for years. In the end, the decisions publishers make today on AI content rights will determine the financial health and future of journalism itself. Every organization needs to take a hard look at its own books and market position before deciding whether to take on a costly legal battle or pursue strategic partnerships and try to find a place in this new AI-driven world.