Privacy Concerns With Gemini-Powered Siri? This Is What Apple and Google Say

Apple and Google announced a major multi-year artificial intelligence (AI) deal on Monday. The partnership will see the Mountain View-based tech giant offering its Gemini tech stack to the iPhone maker, which will then integrate it with its Apple Intelligence. Most notably, Gemini will also become the brain of the advanced Siri, which the company has been delaying for almost two years. However, the partnership also raises privacy concerns about whether Google will be able to access and train its AI models using the conversations with Siri.

Apple-Google Deal Raises Privacy Questions

When it comes to ensuring the data privacy of users, Apple and Google exist on opposite ends of the spectrum. The iPhone maker has implemented strong data protection measures, ensuring most of the identifiable user information never leaves the device. On the other hand, the Search giant has publicly admitted to training its AI models on publicly available web data. The conversations with Gemini are also saved in Gemini Apps Activity for 18 months by default, and can be used to improve its models unless the user turns it off manually.

So, once Gemini begins powering the AI capabilities in Apple Intelligence, and more specifically in Siri, it will be important to see how the iPhone maker plans to ensure the data privacy of users. In the joint statement with Google, the company did address this in a single sentence, saying, “Apple Intelligence will continue to run on Apple devices and Private Cloud Compute, while maintaining Apple's industry-leading privacy standards.”

But there is no elaboration on the implementation or transparency on what users should expect. However, if Apple's claim is taken at face value, it is likely that Google is developing a custom Gemini model which will be handed over to the iPhone maker. The same model will then run on-device as well as via Private Cloud Compute servers. The Pixel smartphones do something similar, with a custom Gemini nano model offering some AI features directly on-device.

For a complete understanding of the implementation architecture, we will have to wait till the company releases details regarding the same. At the earliest, this will be done when the advanced Siri is released to users, which is expected to be in spring this year.