Gemini Might Be Powering iOS 18’s New AI Features

It’s no secret Apple is quite late to the AI party. Ever since the generative AI boon of late 2022 and early 2023, in which every corner of the tech world was doing something with artificial intelligence, the world has wondered what Apple would do with the technology. So far, not much.

However, all rumors point to Apple’s big software updates, like iOS 18, to introduce the company’s first major forays into artificial intelligence. It wouldn’t be a surprise to update your iPhone this year and find a version of Siri that resembles ChatGPT, or to see a helpful AI assistant when using an Apple app like Pages or Keynote. It’s nothing particularly groundbreaking, but it would put Apple in line with other companies that are integrating AI into their platforms.

Of course, we all expected Apple to power these features with its own proprietary, large language models (LLMs). But apparently, Apple might be calling in reinforcements.

According to Bloomberg, Apple is currently in talks with Google to use Gemini to power the iPhone’s upcoming AI features. While the idea of Apple using another company’s AI tech is certainly a surprise—let alone a competitor as big as Google—Gemini wouldn’t be Apple’s only AI engine. Apple would reportedly still use its own LLMs for on-device generative AI: Anything your iPhone could process on its own, that’d be Apple’s doing. However, any AI tasks that require outsourcing to the cloud would be handed over to Gemini.

This is not necessarily a great look for Apple. Look, they wouldn’t be the first company to do this: Samsung partnered with Google to power its Galaxy AI features, and Google could strike similar partnerships with other companies. But this is Apple, after all. Sure, they partner with companies like Google for certain features and advancements (Google has paid big bucks to be the default search engine on Safari for years) but I can’t imagine they’d want to use competitor AI tech unless they absolutely had to. I can see the advertisements touting Apple’s proprietary AI as not just the “most advanced” on the market, but also the most secure. It’s a tougher sell when half your features are being sent to Google’s servers and processed with Google’s AI.

That said, if Apple really is as far behind in the generative AI space as some think they are, they may need to commit to a partnership like this. Investors seem to like it: Apple shares were up 2.5% after the news, while Alphabet stock jumped 6%. Apple gets the AI tech it needs to implement features other companies already have, while Google makes a bunch of money in the process. Not to mention, the move would be increased competition against OpenAI and Microsoft: Apple and Google would run on Gemini, while Microsoft’s Copilot runs on ChatGPT.

Of course, nothing is set in stone, and this is only based on reports. Apple has yet to really announce anything definitive at all in the AI space as far as features go. Apple researchers have continued to publish their work on AI, such as the company’s MM1 AI model, open-source image editor, and image animator. And while some or all of these developments may make their way into actual iOS or macOS features, we have yet to see them.