Google reportedly paid an eye-popping $2.7 billion to lure back a former employee who quit three years ago when the search giant refused to release the chatbot he had developed.
Noam Shazeer, 48, pocketed hundreds of millions of dollars from the deal to license his own artificial intelligence technology and return as a vice president, the Wall Street Journal reported.
Google co-founder Sergey Brin played a key role in bringing back Shazeer. Brin said at a recent conference said that the company had been too timid when it came to releasing AI applications, the Journal noted.
Google now plans to develop and launch AI technology as soon as possible, Brin reportedly said.
The deal with Shazeer followed similar moves by Microsoft and Amazon, the Journal said, and marked the latest chapter in an ongoing debate over whether Big Tech is spending too much on AI, which both proponents and critics predict has the potential to change life as we know it.
"Noam is clearly a great person in that pace," Stanford Artificial Intelligence Laboratory director Christopher Manning told the Journal. But "is he 20 times as good as other people?" he asked.
Google employees who work on AI said they didn't know what the company would do with the technology licensed from Shazeer's company, Character.AI, and the deal is widely seen as a way to get him to work for Google again, according to the Journal.
Google declined to make Shazeer available for an interview, and he didn't respond to requests for comment, the Journal said.
Shazeer was hired by Google in 2000, and his first major project reportedly involved improving the system that corrects spelling errors entered into the company's now-dominant search engine.
Shazeer failed in his first attempt to develop AI with human-level intelligence, but then-Google CEO Eric Schmidt was bullish on his chances during a 2015 talk at Stanford University, according to the Journal.
"If there's anybody I can think of in the world who's likely to do it, it's going to be him," Schmidt said.
Two years later, Shazeer and seven other Google researchers published a paper that detailed a computer system to predict the next worked in a sequence when prompted by humans.
He then teamed up with colleague Daniel DeFreitas to build a chatbot named Meena, which he predicted could replace Google's search engine and generate trillions of dollars in revenue, the Journal said, citing people familiar with a memo Shazeer wrote about the project.
But Google executives refused to release Meena due to concerns about safety and fairness, the Journal said, and Shazeer and De Freitas quit the company in 2021 to launch Character.
The company struggled to keep up with the cost of developing its various chatbots without a steady stream of revenue, and Shazeer spoke with Facebook owner Meta Platforms and other potential buyers earlier this year, the Journal said.
In a blog post last month that announced the Google deal, Character said that the "landscape has shifted" in the AI business since its 2002 founding, and that it saw an "advantage" in using other AI programs "alongside our own."
In addition to Shazeer, De Freitas and "certain members of our research team" also left to join Google, with general counsel Dominic Perella taking over as interim CEO, the company said.
A Character spokesperson told the Journal that it had more than 20 million active monthly users and was on track to build a consumer-oriented business.
