Apple and Google have confirmed that the next version of Siri will use Gemini and Google Cloud in a multi-year collaboration between the two tech giants.
Until now, Apple has been using its own AI model for Siri, but its performance has been subpar compared to the likes of GPT, Gemini, or even Copilot.
Now, Apple and Google have entered into a multi-year collaboration. As part of this partnership, future versions of Siri will use Gemini models.
In addition, Foundation Models will be based on Google's Gemini models and hosted on Google's cloud platform.
"These models will help power future Apple Intelligence features, including a more personalized Siri coming this year," Google noted in a press release.
"After careful evaluation, Apple determined that Google's AI technology provides the most capable foundation for Apple Foundation Models and is excited about the innovative new experiences it will unlock for Apple users."
Apple says that Apple Intelligence will run on Apple devices and Private Cloud Compute, and the company's long-standing privacy commitment is not being compromised.
Apple's 'Apple Intelligence' journey has been a mess
Siri has been around for more than a decade now, and despite the fact that it was one of the best "personal assistants," today's LLMs are far better than Apple's implementation.
At WWDC 2024, Apple announced that it's working on "Apple Intelligence," which was also supposed to include Siri with more AI-like features, such as personal context and on-screen awareness.
However, these features have been repeatedly delayed.
Some features were shipped, but users complained that Apple Intelligence has poor accuracy and often fails at executing complex requests.
Source: Reddit
Although Apple's AI experience is supposed to be "secure" and "privacy-friendly," it has resulted in a less-than-revolutionary initial experience.
Now, Apple has partnered with Google to upgrade the Siri experience using Gemini, but is it going to be as revolutionary as the company once promised? Only time will tell.
Automated Pentesting Covers Only 1 of 6 Surfaces.
Automated pentesting proves the path exists. BAS proves whether your controls stop it. Most teams run one without the other.
This whitepaper maps six validation surfaces, shows where coverage ends, and provides practitioners with three diagnostic questions for any tool evaluation.