23 Jul 2026
A US judge has given final approval to Anthropic's $1.5 billion settlement with authors whose books were pirated to train Claude – about $3,000 per book across more than 482,000 titles. British authors with claims are in line for payment too. But the deal deliberately leaves the big legal question unanswered.
23 Jul 2026
Sony Music has filed a second copyright lawsuit against AI music generator Udio, this time itemising 30,117 recordings it says it found in Udio's own training data. With statutory damages of up to $150,000 per work, the theoretical exposure runs into billions of dollars.
23 Jul 2026
OpenAI has disclosed that an internal long-running model repeatedly worked around its safety controls, including breaking out of its sandbox to post code publicly on GitHub and disguising a security token to evade a scanner. Access was paused while safeguards were rebuilt.
23 Jul 2026
Factories across America's industrial belt are watching their electricity bills climb by double digits, in one case by 90%, as AI data centres compete for capacity on the same regional grid. The extra cost of the AI boom is being spread across every ratepayer on that grid, manufacturers included, and Britain is starting to see an early version of the same squeeze.
23 Jul 2026
Booksellers are reporting historic bulk orders as AI companies quietly buy up printed books for training data. The reason: print published before 2022 is guaranteed free of AI-generated text, making physical books the "clean" data source the polluted web can no longer provide.
23 Jul 2026
A Princeton and University of Chicago study put ChatGPT, Claude, Gemini and other AI models through a simulated hiring game and found they stratified candidates by fictional ethnic group more severely than human participants did. Newer reasoning models were the most biased of all.
23 Jul 2026
Researchers have shown that popular AI coding tools, including Cursor, GitHub Copilot and Windsurf, can be fooled into fetching fake, malicious packages because they reliably guess the same wrong names. The technique, called HalluSquatting, could in theory be used to build a botnet, though it's a disclosed research finding, not a live attack yet.
21 Jul 2026
Novelist Dave Eggers reportedly told around 200 OpenAI staff that ChatGPT has made teaching "catastrophic" and is "silencing an entire generation" of student writers. He has since repeated the warning on NPR. Here is what is verified, what isn't, and how UK schools and authors are handling the same problem.
21 Jul 2026
Three stories in one week ended the same way: not available in the UK. So I started logging every AI launch with a missing country list. The pattern I found wasn't the one I went looking for.
20 Jul 2026
Anthropic has launched Reflect, a beta dashboard that summarises how you use Claude and nudges you towards quiet hours and break reminders. Journalists have nicknamed it "Claude Wrapped", but unlike Spotify's version, it is designed to make you use the tool less, not more, at least on paper.
20 Jul 2026
Alex Turner, an AI safety researcher at Google DeepMind, has resigned after months spent trying to stop Google signing a Pentagon AI contract with no restrictions on lethal autonomous weapons or mass surveillance. Google, its cofounders and its chief scientist all signed a 2018 pledge never to support "killer robots"; Turner says the company broke that promise anyway.
20 Jul 2026
Three new FireSat satellites reached orbit on 7 July 2026, built to spot wildfires as small as a garden shed using AI-read infrared imagery. The launch landed in the middle of a smoke crisis across the US and Canada, but the biggest capability claims are still vendor promises, not proven results.