Apple has poached dozens of artificial intelligence experts from Google and set up a secret European lab in Zurich as the tech giant builds a team to compete against rivals in developing new AI models and products.
The $2.7 trillion company has been on a recruiting spree in recent years to expand its global AI and machine learning team, according to a Financial Times analysis of hundreds of LinkedIn profiles as well as public job postings and research papers.
The iPhone maker has particularly targeted Google employees and has recruited at least 36 specialists from its competitor since John Giannandrea was poached in 2018.
While the majority of Apple’s AI team works in offices in California and Seattle, the tech giant has also expanded a significant outpost in Zurich.
Professor Luc Van Gool of the Swiss university ETH Zurich said Apple’s acquisition of two local AI start-ups – virtual reality group FaceShift and image recognition company Fashwell – led Apple to set up a research lab in the city called Vision Lab build.
Employees based in Zurich have been involved in researching Apple’s underlying technology, which powers products such as OpenAI’s ChatGPT chatbot. Her work has focused on increasingly advanced AI models that integrate text and visual input to generate responses to queries.
The company has advertised positions in the field of generative AI at two locations in Zurich, one of which is particularly little known. A neighbor told the FT that he didn’t even know the office existed. Apple did not respond to requests for comment.
Apple is typically tight-lipped about its AI plans, even as major tech rivals like Microsoft, Google and Amazon tout billions of dollars in investments in the cutting-edge technology.
Its shares have fallen since the start of the year, increasing pressure on the tech giant to announce groundbreaking AI features that could boost device sales.
Industry insiders suspect Apple is focused on bringing generative AI to its mobile devices, a breakthrough that would allow AI chatbots and apps to run on the phone’s own hardware and software rather than being powered by cloud services in data centers .
CEO Tim Cook told analysts that Apple has “explored a broad range of AI technologies” and has invested and innovated “responsibly” in the new technology.
However, the technology group has been developing AI products, such as its voice assistant Siri, for more than a decade. The company has long been aware of the potential of “neural networks” – a form of AI inspired by the way neurons in the human brain interact and a technology that underlies groundbreaking products like ChatGPT.
Chuck Wooters, an expert in conversational AI and LLMs who joined Apple in December 2013 and worked on Siri for nearly two years, said: “During my time there, one of the pushes in the Siri group was to move to one.” neural architecture for speech recognition. Even back then, before large language models came along, they were big proponents of neural networks.”
This interest appears to have led Apple to researchers who were the driving force behind neural networks that power AI models.
In 2016, Apple acquired Perceptual Machines, a company founded by Ruslan Salakhutdinov and two of his students at Carnegie Mellon University that worked on generative AI-powered image recognition.
“At that time they were looking for a whole bunch of researchers and trying to build the infrastructure to train these models,” Salakhutdinov told the Financial Times.
A key figure in the history of neural networks, Salakhutdinov 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 worked as director of AI research at Apple until returning to Carnegie Mellon University in 2020.
Apple’s top AI team now consists of former key Google figures, including Giannandrea, who previously led Google Brain, the search company’s AI lab that has since merged with DeepMind.
Samy Bengio, senior director of AI and ML research, was formerly one of Google’s top AI scientists. Ruoming Pang, who leads Apple’s Foundation Models team working on large language models, previously led Google’s AI speech recognition research.
The company also once hired Ian Goodfellow, another deep learning pioneer, but he returned to Google in 2022 and protested Apple’s return to work.
Six former Google employees hired in the past two years were listed among the authors of a major research paper published in March in which Apple announced it had developed a family of AI models called “MM1.” , which use text and visual input to generate answers.
Apple has also bought about two dozen AI startups in the last 10 years that focus on applying AI reasoning to image and video recognition, data processing, search, and music content curation.
According to their LinkedIn profiles, the founders of Musicmetric, Emotient, Silk Labs, PullString, CamerAI, Fashwell, Spectral Edge, Inductiv Inc, Vilynx, AI Music and WaveOne still work at Apple.
Salakhutdinov said Apple has focused on “doing as much as possible on the device,” which will bring with it the need for more powerful chips with so-called dynamic random access memory (DRAM) that can process the enormous amounts of data needed to power it AI models are required.
“The next big thing will be ‘AI smartphones’ – and these will require a lot more dram,” said Sumit Sadana, executive vice president and chief business officer of Micron Technology, one of Apple’s chip suppliers.
Sadana added: “The average smartphone memory chip today has about eight gigabytes of memory, but LLMs require at least 12 GB to run.”
Salakhutdinov said another reason for Apple’s slow AI adoption is the tendency of language models to produce incorrect or problematic answers. “I think they’re just a little more cautious because they can’t release something that they can’t fully control,” he added.
Apple’s push into generative AI capabilities could be first seen at the company’s Worldwide Developers Conference in June.
Erik Woodring, an analyst at Morgan Stanley, said the next iPhone “could become much more of a voice-activated, intelligent personal assistant, controlled by an improved Siri, and could, for example, interact with all the apps on your phone using voice control.”
He added: “What we’re looking for at WWDC are previews of one or two AI features that can be game-changers for the average consumer.”