Apple has poached dozens of artificial intelligence experts from Google and created a “secret European laboratory” in Zurich to house a new team responsible for creating new AI models and products, on a paid basis. Financial Times report.
Based on an analysis of LinkedIn profiles carried out by FTApple has hired at least 36 specialists from Google since 2018, when the company poached John Giannandrea to become its top AI manager.
Apple’s main AI team works in California and Seattle, but the company recently expanded its offices dedicated to AI work in Zurich, Switzerland. Apple’s acquisition of local AI startups FaceShift (VR) and Fashwell (image recognition) reportedly influenced its decision to build a secret research lab known as the “Vision Lab” in the city.
According to the report, employees based in the lab were involved in Apple’s research into the underlying technology that powers OpenAI’s ChatGPT chatbot and similar products based on extended language models (LLM). The focus has been on building more advanced AI models that integrate textual and visual input to produce query answers.
The report suggests that Apple’s recent work on LLMs is a natural outgrowth of the company’s work on Siri over the past decade:
The company has long been aware of the potential of “neural networks” – a form of AI inspired by the way neurons interact in the human brain and a technology that underpins revolutionary products such as ChatGPT.
Chuck Wooters, a conversational AI and LLM expert who joined Apple in December 2013 and worked on Siri for almost two years, said: “While I was there, one of the pushes that was happening in the Siri group was to move to a neural architecture for voice recognition. Even back then, before big language models took off, they were big advocates of neural networks.
Currently, Apple’s core AI group includes notable former Google members such as Giannandrea, former head of Google Brain, who is now part of DeepMind. Samy Bengio, now senior director of AI and ML research at Apple, was also previously a leading AI scientist at Google. The same goes for Ruoming Pang, who leads Apple’s Foundation Models team that focuses on large language models. Pang previously led AI speech recognition research at Google.
In 2016, Apple acquired Perceptual Machines, a company that worked on AI-based generative image detection, founded by Ruslan Salakhutdinov of Carnegie Mellon University. Salakhutdinov is considered a key figure in the history of neural networks and studied at the University of Toronto under the technology’s “godfather” Geoffrey Hinton, who left Google last year over concerns about the dangers of generative AI.
Salakhutdinov said FT One of the reasons for Apple’s slow deployment of AI was the tendency of language models to provide incorrect or problematic answers: “I think they’re just a little more cautious because they can’t publish something they can’t fully control,” he said. .
iOS 18 is rumored to include new generative AI features for Siri, Spotlight, Shortcuts, Apple Music, Messages, Health, Keynote, Numbers, Pages, and other apps. These features are expected to be powered by Apple’s in-device LLM, although Apple has also reportedly discussed partnerships with Google, OpenAI and Baidu.
A first look at Apple’s planned AI features is expected to arrive in just over a month, with iOS 18 expected to debut at the Worldwide Developers Conference starting June 10.
