Google employees demand stronger whistleblower protections for AI researchers

On March 8, a group of Google employees published an open letter demanding that state and national legislatures strengthen whistleblower protections for artificial intelligence (AI) researchers and other technology workers. The demand comes in the wake of Google recent shots of the two co-leaders of the Ethical AI research team.

The open letter states that “[a]s others have writing, the existing legal infrastructure for whistleblowing companies developing technologies is totally insufficient. Researchers and other tech workers need protections that allow them to call out harmful technology when they see it, and whistleblower protection can be a powerful tool to guard against the worst abuses of the private entities that create these technologies.

Google fired Dr. Margaret Mitchell on February 19, 2021. Dr. Mitchell’s dismissal followed Dr. Timnit Gebru’s departure from the company in December 2020. Dr. Gebru claims that “Google terminated her after questioning a order not to publish study saying AI that mimics language could harm marginalized populations,” according to Reuters. Dr. Mitchell, co-author of the study with Dr. Gebru, openly criticized Google’s firing of Dr. Gebru.

Dr. Mitchell and Dr. Gebru were co-leads of Google’s ethical artificial intelligence team, which Dr. Mitchell founded. Both researchers were critical of Google even before they left, calling for more diversity in Google’s search staff and fearing the company would censor search that criticized its products.

Dr. Gebru gained notoriety for exposing biases in facial analysis systems. Other examples of racial bias found in AI and algorithms include healthcare algorithms that have been proven to have affected millions of black Americans. In a similar vein, an algorithm was found to have blocked black people from receiving kidney transplants. According to VentureBeat“[i]Over the past year, Google’s Ethical AI team has explored a range of topics, including the need for a culture change in machine learning and an algorithm internal audit framework, issues India-specific algorithmic fairness, application of critical race theory and sociology. , and risks of scale.

The open letter calling for increased whistleblower protection was published by Google Walkout for real change, a group of Google employees formed in 2018 and focused on driving change within the company. The letter states, “Google has chosen to dismantle its work on AI ethics, making it clear that the company will only tolerate research that supports its results. It is a matter of pressing public interest.

In an article published in the UCLA Law Review, Sonia Katyal, co-director of the UC Berkeley Center for Law and Technology, says whistleblower protection is a valuable tool for ensuring AI accountability . “Given the issues of opacity, inscrutability, and the potential role of trade secrecy and copyright law as barriers to disclosure, whistleblowing might be an appropriate avenue to consider in the AI,” she wrote.

“I would say very strongly that the existing law is totally inadequate,” Katyal said. says VentureBeat for an article. “What should concern us is a world where all the most talented researchers love [Gebru] being hired in these places and then effectively muzzled to speak. And when that happens, whistleblower protection becomes essential.


The future must be ethical: #MakeAIEthical | by Google Walkout for real change | March 2021 | Average

James G. Williams