According to Financial TimesApple has lured dozens of AI experts from Google and created a «secret European laboratory» in Zurich. A new team of employees works there to create new AI models and products.
According to the publication’s analysis of LinkedIn profiles, Apple has hired at least 36 specialists from Google since 2018, when John Giannandrea became Apple’s chief AI officer.
Apple’s main AI team is based in California and Seattle, but the company recently expanded its offices to Zurich, Switzerland. It is believed that Apple’s acquisition of local startups FaceShift (VR) and Fashwell (image recognition) influenced its decision to build a secret research lab in the city, known as the Vision Lab.
The lab’s staff participated in Apple’s research into the underlying ChatGPT technology and other large speech models. The researchers focused on developing more advanced artificial intelligence models that incorporate text and visual data to answer queries.
Chuck Wooters, a conversational artificial intelligence expert with a master’s degree in law 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 speech recognition, even then, before the advent of large language models, they were big proponents of neural networks».
Apple’s AI leadership team now includes prominent former Google employees, such as Giannandrea, the former head of Google Brain, who is now part of DeepMind. Sammy Bengio, currently the senior director of artificial intelligence and machine learning research at Apple, was also previously a leading AI scientist at Google. The same goes for Ruoming Pang, who leads Apple’s «Basic Models» team, which focuses on large language models. Pang previously led AI speech recognition research at Google.
In 2016, Apple acquired Perceptual Machines, an AI-based generative image detection company founded by Carnegie Mellon University’s Ruslan Salakhutdinov. Salakhutdinov is said to be a key figure in the history of neural networks and studied at the University of Toronto under the «godfather of» technology Jeffrey Hinton, who left Google last year over concerns about the dangers of generative AI.
Salahutdinov tells the FT that one reason for Apple’s slow AI rollout is the tendency of language models to provide incorrect or problematic answers: «I think they’re just a little bit more cautious because they can’t release something they can’t fully control».
iOS 18 is expected 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 on-device language model, although Apple also announces partnerships with Google, OpenAI, and Baidu.