Going to War Against Deepfakes

Companies are racing to detect videos that have been manipulated with the help of Al-before they cause even more violence and confusion in battle zones.

-  VIVIENNE WALT

AS DAWN BROKE on Oct. 7, air-raid sirens blasted out across Tel Aviv, sending Michael Matias and his girlfriend catapulting out of bed and down to the bomb shelter in their apartment building. Inside, the messages on their cell phones revealed the horror that had set off the alarm: Hamas gunmen were waging mass slaughter on Israelis and seizing hundreds of hostages, less than an hour’s drive from where they huddled in safety.

In his stunned aftershock, Matias was struck by the implications for his tech startup. Late last year he had launched Clarity, an artificial intelligence company focused on detecting deepfakes in election campaigns; he believed such disinformation was an urgent threat to democracies. But with about 1,200 Israelis dead on Oct. 7 and the country at war, that mission would need to wait for now. He scrambled an emergency meeting that morning with Clarity’s team, to trigger an action plan. “We said to each other, ‘Our technology is going to be very meaningful here,’ ” says Matias, Clarity’s CEO.