Google's military work reverses one of its oldest values - and it could jeopardize the company's biggest asset
- Google confirmed that it has a contract with the Department of Defense involving AI technology and drones, but has declined to go into detail.
- Google said the technology is being used for "non-offensive" uses.
- Google has long avoided being part of the military industrial complex, to the point where it seems to have been an unofficial company policy.
- The news that Google was working with the DoD reportedly upset many of the company's employees.
Google crossed a red line on Tuesday.
Actually, Google crossed the line sometime before Tuesday, but it was only with Gizmodo's report about Google's contract providing AI technology to the U.S. Department of Defense that the world learned about the situation.
For anyone who has followed Google for some time, the revelation of its contract with the Pentagon is stunning. That's because Google has long resisted playing any part in the military industrial complex.
It's not explicitly written in any of Google's bylaws, but it's easy enough to recognize.
In the past 10 years of records available on USAspending.gov, there are roughly a dozen public contracts between Google and the Department of Defense.
These contracts are all inconsequential: $10,ooo for access to Google Earth here, or $6,000 for Google's search hardware there. It's less than pocket change for the deep-pocketed Google.
For a company at the leading edge of tech innovation, that's pretty remarkable. Compare that to Microsoft, for instance, which has hundreds of contracts with the DoD during the past ten years, including a $78 million multi-year consulting deal with the Air Force.
Google has since sold Boston Dynamics, and Google is not building killer robots for the military. But the AI contract with the military is clearly a break from tradition - according to Gizmodo, the accidental revelation of the contract within Google set off a firestorm among some employees.
According to the report, the military is using Google's TensorFlow - open source technology that anyone can use to create applications that harness artificial intelligence - to help analyze video footage collected by drones.
The collaboration is part of a pilot project known as Project Maven, and most likely involves Google staffers helping the military figure out how to use TensorFlow and other AI technology. Google is not disclosing the financial terms of the contract.
This is not what most people think the words "non-offensive" means
Google is trying to downplay the revelation as business as usual. The company's official statement stresses that Google has always worked with "government agencies," as if there were no difference between working with the Department of Education and the Department of Defense.
Google also insists that the contract is for "non-offensive" uses.
Still, it's not clear what exactly "non-offensive" really means. Sure, Google technology may not be part of a Skynet-like system that allows drones to automatically fire missiles when Google's AI detects an enemy target.
But Google is using a much narrower definition of "non-offensive" than what most people probably imagine.
Google's AI technology will allow drones to distinguish between a school with students carrying books and a bunker with gun-toting fighters. Yes, it will be up to military analysts to then decide whether or not to strike a target. But to describe Google's role in these scenarios as "non-offensive" stretches credulity.
Google is jeopardizing its biggest asset
This is a big reversal of what appears to have at the very least been an unofficial policy long espoused by the company that brought you Don't be Evil.
It's worth noting that former Google CEO and Executive Chairman Eric Schmidt was appointed to a Pentagon advisory board in March 2016 aimed at bringing Silicon Valley innovation to the military. In December 2017, Schmidt resigned his position as Executive Chairman of Alphabet (though he remains on the board and a technical advisor) so that he could "dive into" the latest science, technology and philanthropy.
Whatever the cause of the change, it's alarming because Google's biggest asset is consumer trust. Specifically, the trust that a company that has accumulated so much personal data about our lives (from our search history, to the places we go using Google Maps) and that has so much power, will remain loyal to the principles on which founders Larry Page and Sergey Brin built the company.
As Google, and its now-parent company Alphabet, expand into more areas of our lives (Waymo for cars, Verily for health), and as the company's founders get older, it's reasonable to wonder if it still represents the values it once did.
To be clear, there's nothing wrong with Google working with the military. Plenty of tech companies do it. It's every company's prerogative to make that decision. But Google made a choice many years ago to not dedicate its technology to warfare.
If it's going to change that, especially as Google becomes a major developer of sophisticated artificial technology, it owes the world and the employees working on that technology an honest explanation.