Apple is in discussions with Google to license Gemini for AI capabilities on iPhones following OpenAI deal.

Apple is reportedly in discussions with Google to utilize its Large Language Model (LLM) Gemini for AI capabilities on the iPhone. This news follows Apple’s recent acquisition of Canadian AI startup DarwinAI, leading up to the anticipated release of iOS 18 with significant AI enhancements for the iPhone.

Apple is negotiating with Google

Bloomberg told people familiar with the matter that the Cupertino-based tech giant is in “active negotiations” with Google to license its Gemini artificial intelligence engine as it seeks to catch up with rivals such as Google and Microsoft. they have several notable AI features in their respective smartphones.

Apple is also said to have held talks with OpenAI to see if it could use its generative AI model to bring AI features to the iPhone this year.

While the terms of the deal or branding have yet to be finalized, Bloomberg reports that any announcement about this major partnership between Apple and Google could be held until June, when the iPhone maker hosts its Worldwide Developers Conference (WWDC).

iOS 18 to bring artificial intelligence to the iPhone

Previous reports have suggested that Apple is banking on iOS 18 to bring AI capabilities to the iPhone as it plays against its rivals. Bloomberg’s Mark Gurman has already hinted at it as one of the “biggest updates” in iPhone history as it aims to support Apple’s artificial intelligence.

Thus, one of the biggest beneficiaries of this AI work is expected to be Siri, Apple’s iPhone assistant, and may be the powerhouse of the LLM. But the company probably won’t stop there. Reports suggest that Apple is looking to incorporate AI into as many apps as possible, including Apple Music, Keynote, Pages and even AppleCare.

To meet these AI demands, the iPhone maker started building about 100 AI servers in 2023, and more may come this year, said Jeff Pu, an analyst at a Hong Kong investment firm. It may also bring in-device AI technology called “edge AI”.

A step forward in artificial intelligence

This big news comes just days after it was revealed that Apple researchers have had a breakthrough in training AI models on both text and images. A research paper on the same was published on March 14. In “MM1: Methods, Analysis & Insights from Multimodal LLM Pre-training”, it shows how using multiple architectures for training data and models can help achieve top-level results with multiple benchmarks.

“We show that large-scale multimodal pre-training using careful caption, interlaced caption, and text-only data is critical to achieve state-of-the-art few-image results with multiple benchmarks.” , the researchers said.