The AI news that matters this week

Siri's new AI gets a memory, but it still isn't switched on in the EU

Apple released iOS 27.2 on 26 October. Siri's AI features can now remember earlier conversations, so you can ask about something it suggested before. They can also draw on more of your own information, such as photos sent in Messages, people invited to Calendar events, voicemails and your Safari history.

Siri AI now works in French, Japanese, Korean, Portuguese and Spanish as well as English, with more languages promised by the end of the year. A new "Excluded Apps" setting lets you choose, in one place, which apps Siri and suggestions can look into.

The catch for Europeans: Siri AI is still not available in the EU. Apple first delayed it in June, pointing to the Digital Markets Act, and says it is working with the EU to find a way forward. The features also need an iPhone 15 Pro or newer.

Sources: The Next Web, Ana Maria Constantin

Microsoft offers voluntary buyouts to about 60,000 staff

Microsoft has offered voluntary exit packages to around 60,000 employees in the United States, Canada, India and the United Kingdom, roughly a quarter of its workforce, according to the Associated Press.

The move comes as tech giants pour record sums into AI data centres. Microsoft's Chief People Officer, Amy Coleman, said the voluntary exit offer is new for the company and that it will learn from it, while helping people build new skills.

Sources: Associated Press, Michelle Chapman (via FOX5 Vegas)

Proton will "re-interview" staff about how they use AI

Proton, the Swiss company behind Proton Mail and Proton VPN, plans to re-interview some employees between 4 and 11 November about how AI has changed their work. A leaked internal email, first reported by 404 Media, says the results could affect roles and even continued employment. Staff were asked to bring concrete examples and results.

Proton says it is not doing layoffs "using AI as an excuse". It says it is short of talent, is hiring, and wants to find out who needs extra support.

Sources: 404 Media · TechRadar, Monica J. White

Anthropic, the company behind Claude, files to go public

Anthropic has filed its paperwork (an S-1) with the US Securities and Exchange Commission to list its shares, under the proposed ticker ANT. Goldman Sachs and Morgan Stanley lead the group of banks handling the offer, and Bloomberg reports that marketing to investors could start as soon as the week of 9 November.

It would be one of the first chances for ordinary investors to buy shares in a leading AI lab directly.

Sources: IPO Scoop

EU AI Act: the strict "high-risk" rules now start in December 2027

A new EU law, the Digital Omnibus on AI (Regulation (EU) 2026/1744, published on 24 July 2026), pushed back the AI Act's rules for "high-risk" uses, such as AI used in hiring, credit and education. They now apply from 2 December 2027 instead of August 2026.

Other parts were not delayed. The transparency rules in Article 50, for example on making it clear when content such as deepfakes is AI-generated, have applied since 2 August 2026.

Sources: EUR-Lex, Regulation (EU) 2026/1744 · ActuIA