Demis Hassabis, Head of Google Deepmind, speaks during a press conference ahead of the Google DeepMind Challenge Match on March 8, 2016 in Seoul.
Jung Yeon-Je | AFP | Getty Images | Getty Images
LONDON – Artificial intelligence researchers don’t like when you ask them to name the best AI laboratories in the world, possibly because it’s so difficult to answer.
There are some obvious competitors. US Big Tech – Google, Facebook, Amazon, Apple, and Microsoft – have set up dedicated AI labs over the past decade. There’s also DeepMind, owned by Google parent company Alphabet, and OpenAI, which counts Elon Musk as a founding investor.
“Wow, I hate that question,” Georgia Tech School Associate Professor of Interactive Computing, Mark Riedl, told CNBC when asked to identify his excellence.
“In terms of reputation, there is a good argument that DeepMind, OpenAI and FAIR (Facebook AI Research]) are the top three,” said Riedl.
AI investor Nathan Benaich, a partner at Air Street Capital, agreed. Google Brain and Microsoft could potentially make it to the top, Benaich said, before adding that he felt that Amazon and IBM are not quite in the same league when it comes to the results and impact of AI research.
Another AI expert, who asked to remain anonymous because he didn’t have permission from his company to speak publicly, told CNBC that DeepMind, OpenAI, and FAIR are likely the top three AI-only research labs in terms of known funding are more patents while IBM is pushing this. “The unknown question is the Baidus and the Tencents of the world,” he said of the Chinese tech giants.
Alphabet gives DeepMind hundreds of millions of dollars to do its work annually, while Microsoft invested $ 1 billion in OpenAI in addition to the $ 1 billion that its founding investors contributed. FAIR’s funding is less clear because Facebook doesn’t break it down.
The potential of AI
One way to measure the impact of an AI laboratory is to look at how much science it publishes at the two largest AI conferences: NeurIPS and ICML.
In 2020, Google had accepted and published 178 articles on NeurIPS, while Microsft had 95, DeepMind 59, Facebook 58 and IBM 38. Amazon had less than 30.
In the same year, Google accepted and published 114 articles at ICML, DeepMind 51, Microsoft 49, Facebook 34, IBM 19 and Amazon 18.
PR versus reality
AI has been hailed as a technology that has the potential to usher in a new industrial revolution and dramatically change the world we live in. But at least for the moment she remains relatively young and “tight” in her skills – an AI that can play chess on a superhuman level doesn’t know how to make an omelette, for example.
DeepMind, OpenAI and FAIR are widely considered to be the top three laboratories, partly due to “strong PR games,” said Riedl.
He believes Microsoft, which does much of its AI work through Microsoft Research, could rightly be included in the top echelons. “For some reason they sometimes fly under the radar,” said Riedl. “Salesforce, Amazon, and IBM all have some really strong research pockets, but they can’t make a big splash here either.”
Riedl said he was “not sure if you could swap a group of researchers from one of these companies with another and make a difference.”
Neil Lawrence, former director of machine learning at Amazon Cambridge, told CNBC that Amazon doesn’t have a large, centralized AI research lab because it is more focused on bringing technology to customers.
“I would say they did this very successfully,” said Lawrence, who is now a professor of machine learning at Cambridge University. “But if you take (academic) publications as a measure, they have no rank.”
Lawrence said that Microsoft Research personally is the research lab he admires most, but “Amazon is really a leader in providing AI … even though there isn’t a (large) research lab.”
He added, “DeepMind, OpenAI and FAIR have definitely dominated the headlines. But it’s interesting how much of the research they publish has traditionally been done in universities.”