The Siri application icon in October 2025. Nikolas Kokovlis/NurPhoto/Getty Images

It’s not a hell-freezes-over scenario, but it’s certainly an eyebrow-raiser.

Apple has quietly confirmed that it will indeed use Google’s Gemini AI as well as Google Cloud services to support its lackluster Siri personal assistant.

The companies said in a joint statement that “Google's Al technology provides the most capable foundation for Apple Foundation Models,” which is both a statement of fact as well as a painful acknowledgement that Apple’s own AI, well, isn’t.

Apple noted that Apple Intelligence features will continue to run “on Apple devices and Private Cloud Compute.” 

The arrangement, a multiyear partnership between two of tech’s biggest bigs, buys Apple more time to get its AI house in order and expands Google’s AI footprint across the global mobile device landscape.

It also echoes the multibillion-dollar arrangement that Apple made with Google in the aughts to be the default web search service for its very first iPhone.

Shares of both companies were up slightly in the wake of the news. —AN

You win some, you lose some. Even if you’re Mark Zuckerberg.

Meta reportedly plans to continue the grand unwinding of its metaverse ambitions by laying off 10% of its 15,000-person Reality Labs division, which is responsible for the company’s virtual reality headsets and Horizon Worlds game.

According to the New York Times, layoffs will be announced this week—as soon as today, even.

Not that Meta plans to spend less money on headcount, mind you. 

The Facebook and Instagram parent has vowed to continue plowing money into AI this year after spending $75 billion on it last year, forecasting bigger capital costs that could outpace revenue growth (and investors’ patience).

Meta has already boosted the budgets of various AI teams, such as its TBD Lab working on superintelligence. Expect some of the money that once flowed to VR headsets to redirect to AI wearables. (Tomato, tomahto.)

It’s not quite the final chapter for the company’s metaverse ambitions, which have been a strategic priority for a decade. But with each passing year, it’s becoming ever more clear that the only “meta” left at Meta might be the company’s name. —AN

Google has removed some of its so-called AI Overviews in the wake of a report showing that some of them provided harmful or misleading medical information.

An investigation published by the Guardian at the very beginning of the year prompted Google’s AI for the “normal range” for “liver blood tests” or “liver function tests” only to receive lots of numbers, little context, and “bogus information,” the news outlet said.

“The summaries could lead to seriously ill patients wrongly thinking they had a normal test result, and not bother to attend follow-up healthcare meetings,” the Guardian wrote.

Google said it doesn’t comment on "individual removals” within its Google Search product but acknowledged that some AI Overviews “miss some context,” leading it to “take action.”

It’s a narrow win. Prompting Google for the same information in rephrased queries again surfaced the problematic AI Overviews, the Guardian said—another sign that the management of inaccurate AI is a cat-and-mouse game without end. —AN