Google has fired Satrajit Chatterjee, a member of its “Brain” team, a deep learning research project, after questioning the validity of a company’s research paper. He and his team reportedly challenged a research paper published by Google in 2020 on using artificial intelligence to make computer chips. The article, published in the scientific journal Nature, claims that “computers can design components for certain chips faster and more efficiently than humans.”
Google is mired in the artificial intelligence (AI) search ethics scandal that has rocked the company for more than a year. She is once again being criticized by the scientific world for firing artificial intelligence researcher Satrajit Chatterjee, following the dismissal of Timnit Gibero in December 2020 and the dismissal of Margaret Mitchell in February 2021. So how did this happen? In fact, in 2020 Google published a study stating that artificial intelligence can be more efficient than humans at manufacturing chips. This may indeed take chip design into a new era.
A Google study provided a way for artificial intelligence to automatically generate parts of a computer chip more efficiently than humans. But Satrajit Chatterjee recently expressed his reservations about this assertion and other conclusions of the study. According to sources familiar with the matter, Chatterjee was allowed to contribute an article opposing these claims after he and other authors expressed reservations. They believe Chatterjee, 43, would have presented his counter-arguments in an internal email, and would also have questioned whether the technology had been fully tested.
However, Chatterjee was fired after the article was rejected by the decision committee and the team had hoped to refer the case to CEO Sundar Pichai and Alphabet’s board of directors. Google did not elaborate on why Chatterjee was fired, but said he had been “terminated because of”. The company also confirmed that the original article in Nature had been “rigorously checked” and peer-reviewed. “We thoroughly reviewed the original Nature paper and stood by the peer-reviewed findings,” said Zubin Kahramani, Vice President of Google Research.
Kahramani also said Chatterjee’s study challenging the allegations “does not meet Google’s standards.” “We also thoroughly reviewed the technical claims for a later submission, and it did not meet our publication criteria,” he added. On the other hand, Chatterjee’s team argued that Google violated its own AI principles by rejecting the article. As for critics, they think: Whether or not the science holds up, this release highlights the ongoing clash between Google’s management and the AI teams that drive many of its projects.
In fact, this isn’t the first time that Google has fired an AI researcher, and Chatterjee’s dismissal has only fueled tensions over academic freedom and diversity that have grown within the company since 2020. We were made aware of the situation when Google fired AI ethicist Timnit Gebru in 2020 for A dispute over an article, followed by two others. It escalated when internet giant Margaret Mitchell was fired in early 2021. As a reason for the dismissal, Google claimed that Mitchell had violated its data privacy policies.
But Mitchell publicly criticized the administration’s decision to fire Gebru. These layoffs led to senior engineers leaving the company and demanding overhaul, not to mention company-wide protests over military contracts. Google investigated Gebru’s dismissal and promised to make changes, but Chatterjee’s dismissal may not ease tensions. Also note that Alex Hanna, an ethical AI researcher at Google, published a letter of resignation earlier this year, describing the “albedo problem” within Google and other tech companies.
Hannah was followed by software engineer Dylan Baker who joined the Gebru Institute for Distributed Artificial Intelligence Research, which launched late last year. On the other hand, other critics believe that the tension between some AI researchers and Google can also be explained by the fact that “Google is very proud of its AI research.” The internet search giant has made its AI research central to the company’s reputation for innovation. Pichai said AI would be more important than electricity or fire. Google has already integrated AI models into its core search product to provide users with better results.
Google is keen to promote algorithms that can outperform humans in areas such as chip design and cancer detection, and its flagship Pixel 6 phones are built around a custom processor driven by artificial intelligence. Any flaw could significantly hamper Google’s business, not to mention tarnish its reputation as a leader in the development of artificial intelligence.
Source: Google research article (PDF)
And you?
What do you think about it?
What do you think of this new layoff within the Google AI team?
See also
The second head of Google’s ethical AI research team has been sacked, amid employee protests over Timnit Gibero’s dismissal in December.
Google chief Sundar Pichai apologizes for firing AI ethics researcher, vows to investigate her departure
More than 1,000 Google employees have signed a letter to hold the company accountable for the sudden dismissal of their colleague Timnit Gebru, a leading AI researcher.