AI Weapons Need Human Safeguards
A former intelligence watchdog warns that the vast funding asymmetry between AI development and oversight is putting the U.S. in danger.
Michael K. Atkinson is the former Inspector General of the Intelligence Community in the Office of the Director of National Intelligence.
Frances Haugen, the former Facebook employee whose urgent concern about the lack of oversight for the company’s artificial intelligence systems have generated headlines and hearings in the U.S., the U.K., and soon the EU, may be the world’s most prominent AI whistleblower, but she will not be the last. Companies could spend nearly $342 billion on AI software, hardware and services in 2021, and such AI spending remains on track to break the $500 billion mark by 2024, according to the latest edition of International Data Corp.’s Worldwide Semiannual Artificial Intelligence Tracker. As AI becomes more pervasive, profitable, and powerful—and until we train AI to blow the whistle on itself—we should ensure that humans retain the ability to speak up about its potential harms.
I am familiar with the challenges and consequences of speaking truth to power. Airing and investigating urgent concerns, the Intelligence Community’s term for serious problems, abuses or deficiencies relating to the funding, administration or operation of intelligence activities, was a critical part of my former position as the government’s top IC watchdog. Although the one that I brought forward related to the Ukraine whistleblower’s complaint, which led to former President Donald Trump’s first impeachment, is widely known, I also shed light on an urgent concern in April 2019 about the lack of government funding for AI oversight.
