This week over 3,000 Google employees—about 3 percent of the company—signed a letter to the company CEO Sundar Pichai demanding that Google end its contract with the Pentagon to develop AI systems for military drones.
“We believe that Google should not be in the business of war,” the letter reads. “Amid growing fears of biased and weaponized AI, Google is already struggling to keep the public’s trust. Building this technology to assist the US government in military surveillance—and potentially lethal outcomes—is not acceptable.”
The Pentagon contract in question is known as Project Maven, which according to a Department of Defense press release from last year aims to “deploy computer algorithm to war zones.”
“Project Maven focuses on computer vision—an aspect of machine learning and deep learning—that autonomously extracts objects of interest from moving and still imagery,” the DoD release reads, claiming that the US has found itself in an “AI arms race.”
In March, it was revealed for the first time that Google was using its TensorFlow AI algorithms to analyze massive amounts of video surveillance footage captured by US drones overseas.
The Google letter warns that if the company continues to pursue the Project Maven contract, it will be in bed with less reputable military contractors like Raytheon, well known for its missile development, and Palantir, which was recently criticized for its deployment of predictive policing algorithms in New Orleans.
The employee letter sees Google’s involvement as conflicting with its mission statement of ‘Don’t Be Evil.’ The company’s decision to involve itself in the project is an interesting one, especially considering that its famous AI division, DeepMind, has touted its AI ethics board as a leader in non-malicious use of machine learning for years.
In addition to canceling Google’s involvement with Project Maven, the employee letter demands that the company “draft, publicize, and enforce a clear policy stating that neither Google nor its contractors will ever build warfare technology.”
In a statement issued by Google on Tuesday that did not directly address the employees protest letter, the company said that “any military use of machine learning naturally raises valid concerns” and that Google’s involvement with Project Maven was “specifically scoped to be for non-offensive purposes.” According to the New York Times, the company did not release the actual language of its contract with the military, however.
It’s heartening to see Google’s employees take a stand on the non-military use of the company’s AI technology, even if it’s just a small fraction of the company’s nearly 90,000 employees. Still, it will be difficult for Google to disentangle itself from from its military ties, even if it wanted to. Former Google CEO and current Alphabet board member Eric Schmidt belongs to the Defense Innovation Board, a Pentagon advisory board, as does Google VP Milo Medin.
For now, Google has taken the palliative measure of saying it is involved in a “comprehensive discussion on this important topic” across the company. Whether or not it will actually extricate itself from this lucrative military contract remains to be seen.