According to Bloomberg, Apple is actively negotiating to incorporate Google Gemini or OpenAI’s ChatGPT as the built-in AI assistant on iPhones.
Citing insider sources, the report states that discussions between the two companies involve integrating Google Gemini into iOS 18, slated for release in the fall of this year. This integration aims to offload certain functions such as image generation, short text composition, and queries to the cloud, enhancing the responsiveness to user-typed or voice commands and potentially closing the gap with rival assistant Siri. The report also mentions that Apple is having similar discussions with OpenAI.
However, the specific branding and other details of this collaboration remain undecided. It is unknown whether Apple will announce this partnership at the upcoming Worldwide Developers Conference (WWDC) in June. This development aligns with Apple CEO Tim Cook’s earlier statement that the company would have AI-related announcements within the year.
If this collaboration comes to fruition, it would provide Apple’s iPhones with advanced AI capabilities, potentially boosting sales for the iPhone (or iPad). On the other hand, if over two billion iPhones worldwide adopt Gemini, it would propel Google to surpass OpenAI or Microsoft in the AI market.
Following the news, Google’s stock price surged by 4.4%, while Apple briefly rose by 2.7%, ultimately closing with a modest increase of 0.6%.
Although both parties stand to benefit from this collaboration, Bloomberg and other media outlets speculate that such a partnership would inevitably attract antitrust scrutiny from regulatory authorities. Google’s search engine, as the default on Apple’s Safari (and other browsers), has already drawn attention from the U.S. government and the European Union.
Meanwhile, if Apple indeed negotiates with Google Gemini, it may also incorporate its proprietary technology in the future. Apple is actively developing AI technology, having introduced the Multimodal Large Language Model (MLLM) called Ferret at the end of last year, and recently unveiling another multimodal model family, MM1. Simultaneously, Apple is researching methods for executing large-scale LLMs faster on mobile devices.