Google has fired another AI researcher, shortly after that researcher wrote a paper challenging Google’s work that the company refused to publish.
Sattajit Chatterjee is a machine learning scientist who, along with other researchers, authored a paper challenging earlier works by Google that asserted that computer AI could design computer chips much faster and more efficiently than human engineers.
Google maintains that its original paper has been thoroughly vetted and that the paper by Chatterjee did not meet their standards.
This news comes after Google fired AI ethicist Timnit Gebru in 2020, which led to two other researchers leaving the company shortly thereafter. Then in 2021, Google fired Gebru’s former co-lead and the co-founder of the ethical AI team, Margaret Mitchell. Google claimed that Mitchell was fired for violating code of conduct and security policies, but she was also actively criticizing Google’s firing of Gebru at the time.
According to a tweet from a New York Times reporter, Google said that Chatterjee harassed the original paper’s author, while Chatterjee’s lawyer maintains that he was attempting to maintain scientific integrity at Google. Google has said they fired Chatterjee “with cause.”
Whatever the case, it seems Google has been butting heads with its AI researchers for some time now.
AI is a large part of Google’s future plans, with the company including an AI-focused processor in its flagship Pixel 6 phone. Google’s parent company Alphabet bought DeepMind in 2014 for $400 million and since then has invested billions into the space, including $31.56 billion in research and development last year alone, or 12.3% of its revenue. That means a lot of pressure from investors wanting to see tangible results from AI research. It remains unclear if that played a role in the firing of critical AI researchers.
Google released its end-to-end machine learning platform TensorFlow 2.0 for free in 2019.