Apple is in early discussions with Google about using the tech giant's Gemini AI to power the revamped version of Siri, according to a report by Bloomberg's Mark Gurman. The iPhone maker had first announced its next big update to Siri at WWDC 2024, but it has been delayed by over a year due to engineering setbacks.
The Cupertino-based tech giant has reportedly approached Google to explore building a custom AI model that would serve as the foundation of the new Siri voice assistant next year. Google is also said to have started training a model that could run on Apple's servers.
Prior to discussions with Google, Apple had also explored partnerships with Anthropic and OpenAI to see if either Claude or ChatGPT could serve as the brain for the new Siri.
A decision on which AI model will be used for powering Siri is still several weeks away, and the company hasn't totally ruled out using its internal models as well.
Apple is internally holding a ‘bake-off’ to see which of the two approaches will work the best. The iPhone maker is said to be developing two versions of Siri: one called Linwood, which is powered by its own AI models, and another called Glenwood, which runs on outside AI models.
While Apple executives are said to have been inclined towards Anthropic initially, the financial terms demanded by the AI startup led to Apple bringing in other companies into the mix.
Meanwhile, in the case of a Siri partnership with third-party models, Apple would likely run them via its Private Cloud Compute servers, which use Mac chips for remote AI processing. This would also mean that the new version of Siri won't run on devices.
These discussions are reportedly different from other discussions held by the company to integrate chatbots into Apple Intelligence. The company currently uses ChatGPT to power more complex queries in Apple Intelligence, and Apple has previously confirmed that it is also looking to have a similar partnership with Google.
