Google employees who have worked with timnit gebru have come out in public to challenge the claim that the researcher on the ethics of artificial intelligence has resigned. On Monday, the team posted a letter on Google’s walkout medium account, firmly stating that gebru was fired and did not resign as Jeff Dean, Google’s head of artificial intelligence, said. They also said that the publication censorship policy gebru was supposed to follow had been “unfairly and discriminatively applied.”. < / P > < p > “Dr. gebru’s dismissal was framed as resignation, but in Dr. gebru’s own words, she did not resign,” the letter said. Gebru was asked to meet certain conditions to stay at Google, including transparency around who wants her papers to be withdrawn, the letter said. In the end, the leaders of the ethical AI team said they could not meet these conditions and preemptively accepted her “resignation.”. Her own manager said he was “scared”. < / P > < p > the paper that got gebru fired detailed the potential risks associated with large language processing models, including over reliance on data from rich countries with more Internet access. “As a result, the language generated by AI will be homogeneous, reflecting the practices of the richest countries and communities,” the MIT review of technology wrote. < / P > < p > this research may be a problem for Google, which created a large language model called Bert in 2018, which changed the way it queries search results. Gebru plans to present the paper at the computer science conference in March. On October 7th, she submitted it to Google for internal review. It was approved shortly after midnight on October 8. In a statement, Dean said the team needed two weeks to review. “Unfortunately, this particular paper was only shared one day before the deadline,” he wrote < / P > < p > but gebru’s team refuted this assessment, saying that the review policy was designed to be flexible and that most people did not follow the structure set by Dean. The data collected by the team shows that the vast majority of approvals occur before and 41% of approvals occur after the deadline. “There is no hard and fast rule that papers must actually pass this review within two weeks,” they wrote In an interview with wired, Google executives asked gebru to withdraw her paper or remove her name from it, which she said was like censorship. “You don’t have papers that keep the company happy and don’t point out problems,” she said. “It’s the opposite of being a researcher of that kind.” < / P > < p > gebru sent an email to Google’s “mental women and allies” list service, detailing the prevarication she received on her paper. She also expressed anger at the company’s diversity, fairness and inclusiveness efforts. “Dei [goals and key outcomes], we don’t know where they come from (and will never be met in any case), random discussions,” we need more guidance “rather than” we need to stop the toxic environment that hinders our progress. “Constant fighting and education, your costs, they don’t matter,” she wrote. < / P > < p > her team members share the same frustration, believing that the company’s goal of creating a more diverse and equitable workplace is weak. “They are really trivial demands,” said Alex Hanna, a senior researcher who worked under gebru. < / P > < p > more than 1500 employees signed a Google walkout petition against gebru’s dismissal. “Instead of being accepted by Google as a brilliant and productive contributor, Dr. gebru faces defense, racism, research censorship, and is now sacked in retaliation,” the petition said < / P > < p > “we were frustrated,” Hanna said. “What can you do when someone who is the core and soul of your team is apparently fired for doing ethical research?” Google said the proposed media negotiation rules would put its free services in Australia at “risk”

By ibmwl