Dr. Timnit Gebru, the Staff Research Scientist and Co-Lead of the Ethical AI team at Google is very famous for her remarkable work on AI technologies, especially on Facial recognition. She has voiced out her bold opinions in the favor of the Black community and through her work on algorithmic bias, she proved that in AI, datasets that were used to train algorithms were predominantly white, as a result of which, the facial recognition models had higher error rates in identifying black-skinned people as compared to fair-skinned.
Recently, Dr. Timnit co-authored a research paper that pointed out some flaws in the language models of Google AI, and that resulted in her termination. Google of course does not agree, but deep down, it seems to be her paper that became the root cause of this issue.
Timnit’s paper raised questions about language models like Google’s BERT being too big.
The main questions in her paper were related to the potential risks these large language models possessed and whether the tech companies were doing anything about them or not? Also, the paper pointed out the presence of inherent biases in these language models and questioned about the environmental costs they incurred.
The paper was titled ‘On the Dangers of Stochastic Parrots: Can Language Models Be Too Big?’. However, after an internal review, the company contacted Timnit and asked her to retract the paper. Another demand was to remove the names of Google’s employees from it.
How Timnit Got Fired
In response to Google’s demands for the research paper, Timnit sent an email to the Brain Women and Allies and also asked for the identity of the reviewer to be disclosed to her, which was refused. She stated further conditions for retracting her name from the paper and mentioned that if her conditions were not fulfilled, then she and the company may ‘work on a last date.’ However, it seems that Google was just waiting for something like that because this was followed by an email from Google telling Timnit that her ‘resignation’ has been accepted and is effective immediately!
Following this termination news, Google’s head of AI team, Jeff Dean said that the paper was not up to the mark and did not have enough research work. The co-author of the paper has disputed this claim citing that the paper was based on extensive and thorough research.
Now, almost 1500 of Google’s employees are upset with this termination and have signed a letter for Google in support of Timnit. She feels that she has become a victim of research censorship because she pointed out some of the flaws within the company.
The people and staff supporting Timnit are also demanding the company to have a stronger tolerance for criticism and they should stay true to maintain research integrity at least.
I’m Ava Paul, an experienced news website author with a special focus on the entertainment section. Over the past five years, I have worked in various positions of media and communication at World Stock Market. My experience has given me extensive knowledge in writing, editing, researching and reporting on stories related to the entertainment industry.
