Google's Ethical AI Controversies and Broken Promises

Key Insight

Google's attempts to establish itself as a leader in ethical AI were consistently undermined by internal conflicts and a perceived failure to uphold its own principles, especially after the 2018 Project Maven controversy. Following employee protests over a Pentagon contract using machine learning to improve drone strike accuracy, Google pledged not to renew the deal and published 'AI principles' committing to avoid AI for weapons, harm, surveillance, or reinforcing biases. The company also announced an internal committee and an external advisory council to guide 'responsible development.'

However, the company's commitment to independent oversight quickly faltered. The external advisory council, intended to bolster ethical AI development, faced immediate and severe backlash from thousands of Googlers. Public outcry focused on contentious appointees, including a climate change skeptic and the CEO of a drone company, leading to resignations and the council's complete disbandment barely one week after its formation. This incident highlighted a significant gap between Google's public image of ethical responsibility and its practical implementation, generating cynicism among both employees and external observers regarding the sincerity of its ethical promises.

The most damaging ethical controversy involved the controversial firings of Timnit Gebru and Margaret Mitchell, co-heads of Google's ethical AI unit, who were hired for their expertise in identifying AI bias. Gebru, known for her 2018 study showing facial recognition software accurately identified white men 99% of the time but Black women only 35%, was allegedly fired in late 2020 for refusing to remove her name from a scholarly paper critiquing large language models, a move Google attributed to the paper 'didn't meet our bar for publication.' Mitchell was subsequently fired months later for attempting to investigate Gebru's departure, leading to petitions signed by over 2600 and 800 employees, respectively, who condemned the company's actions as retaliation and censorship of crucial ethical research.