No, Google didn’t just admit that Siri will run on its servers

Ever since Apple and Google jointly announced a partnership to base the Siri voice assistant on Gemini tech, questions have been raised about user privacy. Apple has done its part to offer reassurance by insisting that Gemini-powered Siri will continue to run on-device or on the company’s own Private Cloud Compute (PCC) servers, rather than on infrastructure owned by Google.

But doubts remain. At its triumphant Q1 earnings call late last month, CEO Tim Cook responded to a question about the Gemini partnership by saying, in part: “We’ll continue to run on the device, and run in PCC, and maintain our industry-leading privacy standards in doing so. In terms of the arrangement with Google, we’re not releasing the details of that.”

As 9to5Mac points out, that’s worded in such a way that it could mean other Apple Intelligence features will be hosted in the way described, whereas the new version of Siri will be handled in some other way. I don’t think he was saying that (the “arrangement with Google” almost certainly refers to the financial aspect of the question), but it’s one possible, albeit tenuous, interpretation.

Those doubts resurfaced on Wednesday following another earnings call, this time by Google. The company’s CEO, Sundar Pichai, mentioned the partnership in his remarks in the following way:

“I’m pleased that we are collaborating with Apple as [its] preferred cloud provider and to develop the next generation of Apple Foundation Models, based on Gemini technology.”

Hearing Google refer to itself as Apple’s “preferred cloud provider” while discussing the Gemini partnership is undoubtedly alarming, and I don’t blame pundits for drawing negative conclusions. One news site (9to5Mac, link above) interpreted this as Google “suggesting” Gemini-powered Siri will run on its servers; another (AppleInsider) said Apple and Google’s statements were “seemingly contradictory.” But once again, Pichai’s statement contains vague implications rather than a clear meaning, and I don’t think we should take this as a direct contradiction of Apple’s privacy claims–something which would in effect be a declaration of PR war.

Much like Cook’s statement, there are multiple interpretations of what Pichai said. One is that Apple was lying through its teeth and Gemini-powered Siri will run on Google servers in a way that gives Google access to user data. Another is that Google is simply going to lease server hardware to Apple as part of the arrangement. They will be Apple Park servers running Apple software, but the hardware is “provided” by Google. That would likely fulfil Apple’s privacy pledge, even if the optics might not be great. And another is that Pichai was, clumsily and at the worst possible time, referring to the fact that Google already provides extensive cloud services to Apple in other areas, and that this latest partnership is merely one more example of the two tech giants working fruitfully together.

In other words, the whole thing is frustratingly vague. What we have at the moment is a partnership with many details still to be decided, and the two partners have very different priorities in terms of public messaging. Apple wants to reassure its users at every point that their privacy will be safeguarded, while Google just wants everyone to know that its Gemini AI platform is more important and successful than ChatGPT. So the result is a lot of statements that differ in implication but don’t actually say very much.

There is a fourth option, proposed by Bloomberg’s Mark Gurman. This is that Gemini-powered Siri will run, as Apple has consistently said, on-device or on Apple servers, but that another, more radical version of the voice assistant, a chatbot codenamed Campos, will launch later and run on Google servers. That’s going to be a tough sell, assuming Gurman is correct (and he says only that the two companies are discussing the idea). But Apple will at least have some more time to make the case to its users.