Getty Images, which licenses millions of news, sports and stock photographs, asked the Court of Appeal in London on Wednesday, November 11, to revive its copyright case against Stability AI, the British company behind the image generator Stable Diffusion. Getty says Stability used millions of its pictures without permission to teach the program how to make images, work it says was done in the United States. Last November, a High Court judge ruled largely for Stability. Because the finished program does not keep copies of the photographs it learned from, she found, it is not an illegal copy under British law. Getty won only on limited trademark points involving older versions.

The two-day appeal turns on a narrow question with wide reach: can software trained abroad count as an unlawful copy once it is brought into Britain, even if it holds none of the original pictures? Getty’s lawyer told the judges that Stability valued its collection for training but does not think it should have to pay for it. Stability argues that the images were downloaded and stored in data centers outside the country, so British copyright law does not apply. The BPI, which represents Britain’s record labels, has joined the case on Getty’s side. If Stability wins again, photographers, musicians and writers will find it hard to use British courts against AI companies that train their systems in other countries.

Workday, whose software uses artificial intelligence to sort job applications for employers, asked a federal judge in San Francisco on Tuesday, November 10, to break up a group of more than 15,000 job seekers who say its tools rejected them because of their age. The case was brought in 2023 by Derek Mobley, a Black man over 40 who says he was turned down for more than 100 jobs at companies that used Workday to screen applicants. In May 2025, U.S. District Judge Rita Lin allowed applicants aged 40 and older to pursue the age claim together, and by October more than 15,000 had joined. Workday denies that its technology discriminates.

The company’s main argument is that it does not hire anyone. In its filings, Workday says its customers set their own requirements, use its software in different ways and make their own decisions, so the applicants cannot show that its tools treated them all alike. If the judge agrees, each person would have to sue separately, a far harder and costlier path for someone who never learned why software passed them over. Workday also opposes a request to certify wider groups covering race, sex, age and disability, which Judge Lin is scheduled to hear in March. For people sending out applications, the ruling will help decide whether the maker of a screening tool can answer for its results, or only the employer that used it.

Most American patients are comfortable with artificial intelligence helping to decide their diagnosis or treatment, as long as a doctor can change or overrule its decision, according to a Gallup survey published on November 9. Among U.S. adults who had seen a primary care provider in the past year, 56% said they were somewhat or very comfortable with that arrangement. Only 9% felt the same about AI making those decisions with no review by a provider. The pattern held for reading medical scans and choosing medicines. The survey, run with the nonprofit West Health, questioned 5,330 adults in August. Among patients uneasy with AI acting alone, the most chosen reassurance was the right to ask for a provider’s review, picked by just under half.

Doctors read their patients differently. In a separate online survey of 1,007 primary care physicians, their estimate of how many patients would accept AI under a doctor’s override was 20 points below what patients actually said. Physicians are already using the tools: 89% had used AI for at least one work task in the past year, most often to look up information about medicines, and half used it to write clinical notes. Yet only 20% had received formal training in AI, and just 11% of those in independent practices. For patients, the survey suggests, the condition is not whether AI is used but whether a doctor keeps the final word, which requires doctors trained to check it.