The best tech article I’ve read this year so far is Tom Simonite’s longread on Timnit Gebru and how Google forced her out. But it’s more than that single incident. We hear everything from her origins to her relationships with like-minded colleagues who suffered from similar pushback related to ethics in AI and racism.
I won’t quote lots from the piece as I strongly recommend you take some time to absorb it. But this excerpt pinpoints the moment just before Gebru joined Google in 2018:
Some people warned Gebru about joining the company. While she was interviewing, Google employees were pressuring their leaders to abandon a Pentagon contract known as Project Maven, which would use machine learning to analyze military drone surveillance footage. Gebru signed a letter with more than 1,000 other researchers urging the company to withdraw. Her uncomfortable experience at the 2015 Google party in Montreal preyed on her mind, and multiple women who had worked at Google Brain told her that the company was hostile to women and people of color, and resistant to change.
Gebru considered walking away from the job offer, until [Margaret] Mitchell offered to make her colead of the Ethical AI team. They would share the burden and the limelight in hopes that together they could nudge Google in a more conscientious direction. Gebru reasoned that she could stick close to Mitchell and keep her head down. “I thought, OK, I can do my work and be careful who I collaborate with, and try to ignore some of the other things,” she says. “My number one concern was: Can I survive in this environment?”