Members of Google's ethical AI team reportedly complained about harassment and bias years before being fired

• Google's former lead AI ethics experts reportedly raised complaints of harassment years before being fired.
• Timnit Gebru and Margaret Mitchell flagged bullying and misconduct in 2018, Bloomberg reported.
• Google said some of the accounts were inaccurate and that it investigated harassment allegations thoroughly.

Two of Google's most prominent researchers into AI ethics had flagged issues around sexual harassment and bullying long before they were fired from the company, Bloomberg reported.

Timnit Gebru and Margaret Mitchell, once chief AI ethicists at the tech giant, made headlines when they were ousted from the company in December 2020 and February this year respectively.

According to Bloomberg, the pair raised a number of concerns with senior management over the behavior of some colleagues in 2018, years before they were themselves kicked out.

Complimentary Tech Event

Transform talent with learning that works

Capability development is critical for businesses who want to push the envelope of innovation.Discover how business leaders are strategizing around building talent capabilities and empowering employee transformation.Know More

Bloomberg outlined a litany of complaints by the pair to senior figures within Google.

In one instance, Gebru reportedly informed her superiors that a colleague, whose identity remains undisclosed, had previously been accused of sexual harassment at another company. Google said that it subsequently opened an investigation.

According to the report, both Gebru and Mitchell told the firm's AI chief Jeff Dean about this colleague's past behavior. They also discussed their fears of gender disparity among senior employees, including a "pattern of women being excluded and undermined" on the AI research team, and a number of women employees being assigned lower roles than less-qualified men.

The allegations came in mid-2018, just as new details of sexual misconduct allegations against former Android chief Andy Rubin emerged, which prompted a walkout staged by almost 20,000 Googlers worldwide. Rubin has denied all allegations of misconduct against him.

"I did not go into it thinking this is a great place," Gebru said in an interview with Bloomberg.

"There were a number of women who sat me down and talked to me about their experiences with people, their experiences with harassment, their experiences with bullying, their experiences with trying to talk about it and how they were dismissed."

However, Dean reportedly pushed back against the idea there was any systemic misogyny within the team, but subsequently announced a new research project led by the alleged harasser. Dean reportedly fired this person a short time later, in June 2019, citing "leadership issues ".

Gebru and a number of her co-workers are said to have reported other instances of workplace misconduct, bullying, and obstructive behavior among leadership.

In early 2020, around nine months before she was fired, Gebru says she wanted to examine a dataset publicly released by Waymo, Google's sister self-driving vehicle company, to see if there was any difference in the way its AI detected skin color.

Bloomberg reported that the project was obstructed by months of internal "legal haggling", resulting in Gebru and her team abandoning the project.

Waymo didn't comment on the project directly, but a spokesperson told Bloomberg the company uses "a range of sensors and methodologies to reduce the risk of bias in our AI models."

Meanwhile, Mitchell claims she had been denied a promotion while at Google due to "nebulous complaints to HR about her personality."

Google told Bloomberg that it found no evidence that a HR employee had used those words to describe her.

The company pushed back against claims it had ignored allegations of harassment and said some of the reported accounts were inaccurate.

"We investigate any allegations and take firm action against employees who violate our clear workplace policies," a Google spokesperson said in a statement. "Many of these accounts are inaccurate and don't reflect the thoroughness of our processes and the consequences for any violations."

Earlier this month, Alphabet investor Trillium Asset Management called on Google to introduce better protections for whistleblower employees.

The firm, which reportedly owns around $140 million worth of Alphabet stock, filed a shareholder resolution calling on the company to better workers that speak out against their managers.

Are you a current or former Googler with more to share? You can contact this reporter securely using the encrypted messaging app Signal (+447801985586) or email (mcoulter@businessinsider.com). Reach out using a non-work device.