AI Copyright Lawsuits, Settlement Fights & Accountability Gaps

The lawsuits keep piling up, the money keeps flowing in, and nobody seems to be entirely sure who's responsible for any of it.

Two More Newspapers Sue OpenAI — And the Reproduction Problem Is Getting Harder to Ignore

The Seattle Times and Newsday have joined the growing line of news organizations suing OpenAI and Microsoft, claiming their journalism was used to train AI models without permission and that those models now spit out near-verbatim passages from their reporting. This isn't just a legal argument about who owns what — it's a practical one about whether AI companies built their products on someone else's work and are now competing with the people who created it. Courts are slowly drawing lines around what AI can and can't be trained on, and wherever those lines land, they'll affect what these tools can legally do. The cases most likely to move the needle are the ones, like this one, that turn on fair use doctrine — specifically whether scraping publicly available content for commercial training counts as transformative use. That question doesn't have a settled answer yet, and until it does, the legal status of every AI writing and research tool is genuinely uncertain.

https://www.theverge.com/ai-artificial-intelligence/990932/seattle-times-newsday-lawsuit-openai-microsoft

Authors Are Now Fighting Their Own Publishers Over the Anthropic Settlement Money

When Anthropic (the company behind the Claude AI assistant) settled a copyright lawsuit with authors whose books were used as training data, everyone assumed the money would flow to the writers. Instead, publishers and literary agents are apparently claiming a cut, and authors are pushing back hard. This matters because it reveals something the AI copyright story has mostly glossed over: even when creators "win," the institutions between them and the money have their own interests. If you work with freelancers, license content, or depend on a creative community for your industry, the contracts being written right now will matter for a decade.

https://techcrunch.com/2026/09/06/authors-push-back-as-publishers-and-agents-seek-share-of-anthropic-settlement/

A $3.2 Billion AI Data Center Has an Accountability Gap the Size of a Server Farm

A new Ars Technica investigation untangles the corporate structure behind a massive AI data center — multiple companies layered together in ways that make it genuinely unclear who is responsible if something goes wrong, whether that's an environmental violation, a worker safety issue, or a service failure. Data centers are the buildings full of computers that actually do the processing when you hit "generate," and the investigation found the structure here appears designed to distribute liability rather than assign it. That's not an edge case — it's the question at the center of every AI vendor relationship: when your AI vendor has an outage or causes harm, is there actually a responsible party on the other end of the contract?

https://arstechnica.com/ai/2026/09/the-ai-data-center-boom-is-causing-new-accountability-problems/

Apple's Revamped Siri Had a Moment — Then Quietly Faded Into the Background

A Wired writer tried the beta version of Apple's redesigned Siri this summer, found it genuinely impressive, and then... stopped thinking about it entirely as the full release approached. That arc — excited, intrigued, indifferent — is probably more honest than most AI coverage, and it points to something real: a better assistant that people don't habitually reach for is still a product with a usage problem. Apple's challenge isn't building AI that works, it's building AI that earns a place in how people actually move through their day. For anyone building or marketing AI-adjacent products, this is worth sitting with — capability is not the same as adoption, and adoption is not the same as habit.

https://www.wired.com/story/my-brief-summer-fling-with-siri-ai/

Travis Kalanick Wants Back Into Robotaxis

The founder of Uber, who was pushed out of the company he built, is reportedly positioning his new venture Atoms to enter the robotaxi business — self-driving cars that operate as taxis without a human driver. Kalanick has called this his "unfinished business," which is either poignant or ominous depending on your view of how Uber conducted itself the first time around. The robotaxi market is crowded and expensive, with Waymo already operating in multiple cities and every major automaker with an AI ambitions slide deck. Whether Atoms is a serious technical bet or a founder-redemption narrative dressed up as a startup is a question the next 18 months will answer.

https://techcrunch.com/2026/09/06/travis-kalanicks-atoms-might-be-getting-into-the-robotaxi-business/

The real story of this AI moment might not be the technology at all — it's that nobody has figured out who's in charge when things go wrong, and everyone is hoping that question stays unanswered a little longer.

← More daily briefs