Google employees last year organized in protest against a machine learning project with the U.S. military to analyze drone footage. Googlers were successful in halting Project Maven and the company created AI Principles to govern future products. Those 4,000 anonymous Googlers have today been named the 2018 Arms Control Persons of the Year.
The Arms Control Association every year nominates individuals and groups working to advance “effective arms control, nonproliferation, and disarmament solutions.” Public voting opened in December with 1,200 individuals from over 70 countries participating.
This year’s award goes to the anonymous group of 4,000 Google employees that circulated an internal petition asking CEO Sundar Pichai to end the AI project. Many employees at the time were outraged that their work could be used for military purposes with some ultimately leaving the company.
On Google’s part, partnering with the military was a lucrative way to monetize its investment and lead in machine learning. However, the effectiveness and automation made possible by AI could change the future of warfare.
Technological developments that remove or reduce direct human control over lethal weapon systems could change the nature of warfare and undermine global security. Not only do governments need to work harder to develop new rules to mitigate the risks, but researchers at private institutions and tech companies have a responsibility to step-in, when necessary, to ensure their projects are used responsibly.
Project Maven was revealed in March with Google in June releasing AI Principles that prohibit weapons usage. Google Cloud similarly announced that it would not renew its work with the U.S. Department of Defense.
The Arms Control Association notes how “Google employees helped to change company culture and policy for the better and is an example for others to follow.” Other individuals nominated this year include South Korean president Moon Jae-in and various diplomats.
Google provided an update on its AI Principles last year and its work towards creating an internal review structure for projects. Notably, these guidelines led to the company holding work on facial recognition for third-parties through Google Cloud.