An economic analysis by Goldman Sachs examined employment histories of more than 15,000 people who lost jobs between 2016 and 2025. Among workers classified as more likely to have been displaced by AI in 2025, the search for a new job took about one month longer than for other laid-off workers, and the share still without work a year later was around 10 percentage points higher. Axios reported the findings on November 3. The analysis did not establish that AI directly caused each dismissal: the link to AI displacement was estimated using job characteristics and adoption patterns.

On November 3, an Advocate General of the Court of Justice of the European Union said Google's AI assistant Gemini could, in certain circumstances, infringe press publishers' rights when it retrieves news articles and reuses recognisable parts of their wording in answers. The opinion arose in a dispute involving a Hungarian publisher and Google. It is an Advocate General's non-binding opinion, not a final court ruling. The reported facts themselves are distinct from an article's protected expression, and the use of material to train AI models raises different questions from retrieving and reusing articles in responses.

On November 2, Google UK and the Prospect trade union reached an agreement allowing the union to communicate with workers and provide information about membership at workplaces covering roughly 7,000 employees. The agreement includes Google's UK operations and YouTube, but not Google DeepMind. It follows new UK workplace access rights and allows the union to hold sessions at Google offices. The figure does not mean that 7,000 employees have joined a union, or that new pay arrangements or conditions for deploying AI have already been negotiated.

On November 2, the Hong Kong Monetary Authority (HKMA) launched Project SAGE to prepare for financial services in which AI agents may make purchases or payments on behalf of people and businesses. The regulator established an expert group involving financial institutions and technology specialists to explore how agents should be identified, how a person's delegated authority should be checked, how different services should work together, and how consumer protection and liability should be handled. This does not mean that unrestricted AI-operated consumer payments are already available or that new Know Your Agent requirements have become law.

China’s robotics companies are increasingly recording people’s movements and decisions to train machines for real-world tasks. In a report published on November 2, Reuters interviewed a Spirit AI executive who said the company employs around 1,000 people to record household and industrial work. Another developer described 160,000 hours of training material and said robots were handling orders at two Shanghai pharmacies, with people still checking the medicine. Battery maker CATL has also described robots performing wiring tasks on a production line. Deployment figures and success rates include company claims, however, and this is not evidence that a general-purpose robot can independently perform any job.

Since September, Uber and Wayve have offered a limited passenger service in self-driving vehicles in London, with a safety driver still in the front seat. About 24 vehicles from China’s Baidu are also being tested on London roads. Against this backdrop, the mayor called on November 1 for powers to limit future robotaxi numbers, citing congestion, air quality, and the livelihoods of taxi drivers. The request has not become law, and the limited passenger service still uses safety drivers. Plans to launch a fully driverless service and move drivers into roles such as remote supervision remain proposals; those jobs have not yet been created.