Google has locked out one of its top researchers on the ethics of artificial intelligence from its corporate computer network in what may be a prelude to firing her.
The move is another sign that the technology giant is increasingly taking tough action to choke off dissension within its own ranks and that its relationship with its own in-house A.I. ethics experts has continued to deteriorate.
In late November, Google pushed out Timnit Gebru, a prominent Black A.I. scientist who had served as co-lead of Google Research’s A.I. ethics team. Gebru said she was fired. The company has insisted she resigned.
Gebru’s departure drew widespread condemnation from both outside and inside Google, with close to 6,000 people, including 2,700 company employees, signing an open letter supporting her. In response, Sundar Pichai, the chief executive officer of Google’s parent company Alphabet, issued a company-wide email in early December saying the company would investigate what happened with Gebru, and even acknowledging the company had “progress we still need to make” in improving racial and gender equity within the company.
The latest incident involves Margaret Mitchell, who had served as co-lead of Google Research’s A.I. ethics unit with Gebru, and had continued in that role following Gebru’s ouster. Gebru tweeted yesterday that Mitchell had her access to the Google’s corporate network cut off, and that the company had told Mitchell she would not have access to it again for “at least a few days.”
Mitchell did not respond to a request to comment for this story. The company said in a statement that its systems had automatically locked Mitchell out of the corporate network after “our systems detected that an account had exfiltrated thousands of files and shared them with multiple external accounts. We explained this to the employee earlier today. We are actively investigating this matter as part of standard procedures to gather additional details.”
Diversity record in the spotlight
Mitchell has been a steadfast defender of her former colleague Gebru, publicly criticizing both Google’s treatment of her, and the company’s poor record on diversity and inclusion.
Gebru’s ouster came after she and several colleagues attempted to publish a research paper that called into question the ethics of natural language processing A.I. software that Google has helped develop. After the company asked Gebru to withdraw the paper from publication, she also wrote to an internal company listserv, criticizing the company for paying lip service to racial and gender equity while acting to silence her and other underrepresented employees.
Since then, April Curley, who spent six years at Google as a diversity recruiter, also posted on Twitter that she’d been fired for voicing concerns internally that Google wasn’t doing enough to improve its poor track record on race and gender.
Curley’s posts lead HBCU20x20, an organization that pairs students from Historically Black Colleges and Universities with businesses, to cancel an agreement to work with Google.
Pichai is due to meet representatives from a number of these colleges and universities on Jan. 29 to discuss how the company can improve its record on hiring Black employees.
The clash with the head of its own A.I. ethics team comes at a time when Google is also struggling to deal with a wave of employee activism over everything from its work with the U.S. military to climate change to free speech.
Earlier this month, a group of about 200 workers for Google and its parent company, Alphabet, announced that they had formed a union with assistance from the Communication Workers of America Union.
