Welcome to the wild west of AI copyright lawsuits — where tech giants argue their bots are just really enthusiastic readers, and authors argue they're being robbed blind.

Two major lawsuits against Meta and Anthropic both resulted in key victories for Big Tech. But that doesn’t mean the companies are off the hook.

The core issue in both cases? Meta and Anthropic trained their LLMs on copyrighted material — in some instances using pirated copies, and in Anthropic’s case, later purchasing books in an attempt to legitimize the training process.

Training AI on copyrighted books? Courts are saying: maybe OK under “fair use” at least when the output is deemed transformative. Current copyright law allows for use of copyrighted material in a way that adds new meaning, purpose, or expression, rather than simply copying it.

Getting those books from pirate sites? Still illegal. Even if you later buy them.

Anthropic’s whole brand is built on trust and safety. They’re the calm, buttoned-up cousin in the AI family, so finding out they were engaged in the same pirated data-scraping is surprising.

To their credit, Anthropic’s own team raised red flags about the risks of using pirated books. In response, they brought in the former head of Google Books to help steer things in a more “legitimate” direction: they bought physical copies en masse and started a book-sacrificing and scanning operation. Points for effort?

LLMs need huge amounts of high-quality text to learn, and books are some of the best examples of how humans think and write. But as companies run out of training data and into legal problems, the focus has shifted towards synthetic data, where AI learns from content generated by other AIs.

It might be the next big boost for model performance—and a way to sidestep the copyright minefield. But whether it’s as good as real human writing is still up for debate.

Been reading about AI in the Neural Net and wondering how to actually use it at work? This program’s got you covered.

Apply AI at work with step-by-step guidance from Columbia Business School faculty.

Learn to:

No technical background required.

Enroll today in the AI for Business & Finance Certificate Program from Columbia Business School Exec and Wall Street Prep.

P.S. Save $300 with code SAVE300

Save Your Seat

Another day, another AI glow-up! Claude’s not just chatting anymore — it’s building apps with you.

With Anthropic’s latest update, you can create and share interactive AI apps entirely within the Claude ecosystem — no third-party tools, no code, just human-to-LLM conversation.

It makes sense that Anthropic is stepping into the vibe coding arena, since Claude is already widely regarded as the best LLM for writing and understanding code. If you're using one of the trendy no-code tools, there's a good chance Claude is doing the heavy lifting.

Get the details here.

Salesforce is leaning heavily into AI, with CEO Marc Benioff saying it now handles 30–50% of the company’s work. He calls it a “digital labor revolution,” noting their AI reaches 93% accuracy thanks to rich data and metadata. The shift has already led to over 1,000 job cuts as the company restructures around automation. Other tech firms like Klarna and Amazon are following suit, using AI to streamline operations and reduce headcount.

Kalshi’s wild NBA Finals ad, featuring a bride fleeing the cops in a golf cart and a guy lounging in a pool of eggs, was created in just two days for under $2,000 using AI tools like Google’s Veo 3 and Gemini. The result? A viral hit with millions of views and zero film crews. Marketers and experts say this could be the future of advertising, especially for smaller brands, as lower production costs enable rapid experimentation.

Florida-based footwear startup Syntilay is planning to be the first to design a shoe in space, using AI agents aboard a satellite in low Earth orbit. The design will beam back to Earth, where it’ll be 3D printed and sold as a limited, tokenized collection as soon as 2026. It’s part of a larger commercial mission, featuring crypto mining and AI computing, set to launch on a SpaceX Falcon 9. The goal? Stand out in the sneaker game by finding “white space” in actual space.

For when you want business news that pairs well with your daily dose of AI.

Every day, 3.5 million readers turn to 1440 for their factual news. We sift through 100+ sources to bring you a complete summary of politics, global events, business, and culture, all in a brief 5-minute email. Enjoy an impartial news experience.

Join for free today!

Dine Brands, the parent of Applebee’s, IHOP, and Fuzzy’s Taco Shop, is rolling out AI to streamline operations and boost sales across its 3,500+ restaurants. Here’s what they’re serving up:

Most of the Dine Brands AI work is done with third-party vendors as they join the likes of McDonald’s, Taco Bell, and Pizza Hut in betting on automation to cut costs and improve efficiency.

But it’s not all upside—AI ordering can be glitchy, and Wendy’s faced backlash over dynamic pricing tests. While full robot takeovers aren’t here yet, Dine Brands CIO is bullish: “AI is moving quickly, and I believe it’s going to be embedded in everything we do.”