Big Tech juggernaut Google is one of the top companies for artificial intelligence research. However, the company is in hot water again after yet another Google AI researcher has been reportedly fired for questioning the company's study findings.
Google AI researcher fired… again
Reported by The New York Times, Google Brain — the company's deep learning AI team — fired a high-profile researcher back in March. The outlet explained that the researcher had his contract terminated after questioning a study on AI designed computer chips.
Google AI researcher Satrajit Chatterjee reportedly questioned the study after it claimed AI could design new computer chips faster and more efficiently than humans. Chatterjee is said to have questioned the reliability of the research after the claims were published in Nature, a high-end science journal.
Shortly after, Chatterjee was fired. However, Google’s official termination statement for the researcher doesn't mention anything about the incident. Instead, officially, the researcher was simply “terminated with cause”.
While Google would not tell NYT what the artificial intelligence researcher complained about, the company instead defended its published study.
“We thoroughly vetted the original Nature paper and stand by the peer-reviewed results,” said Google Research VP Zoubin Ghahramani. “We also rigorously investigated the technical claims of a subsequent submission, and it did not meet our standards for publication.”
Google’s Ethical AI issues
Google’s firing of Satrajit Chatterjee is far from the first high-profile termination at the company. In fact, in just two years, this is the third high-profile Google AI researcher to be pushed out of the big tech organisation.
Infamously, Google fired AI ethics researcher Timnit Gebru for bringing attention to dangerous aspects of Google’s AI software. Gebru explained that not enough effort was being put into combating dataset biases which could lead to dangerous outcomes. Afterwards, co-worker Margaret Mitchell was also fired.
This precedent that Google has for firing employees that question findings is dangerous for AI researchers. If one of the top AI companies doesn't let anyone discuss and refute studies, issues that exist will never get fixed, only paved over.