Twitter Releases Ban on Harmful Deepfakes

Yesterday, Twitter released new rules for posting deepfakes on its platform: don't. Remember, deepfakes are persuasive AI-generated media depicting a real person's likeness.

"You may not deceptively share synthetic or manipulated media that are likely to cause harm," Twitter declared. For other deepfakes, Twitter may slap a label on them, reduce the visibility of the tweet, or add more context.

• These rules leave room for interpretation, and ultimately the enforcement ball is in Twitter's court.

Around the deepfake world: Jigsaw, Alphabet's think tank-y subsidiary, revealed a machine learning deepfake detection tool yesterday. A few days earlier, YouTube emphasized it won't allow election-related deepfake videos this year. And Reddit and Facebook banned most deepfakes last month (though FB's measure has critics).

Zoom out: Monitoring deepfakes is a top priority for platform operators, who largely self-police in the U.S. And given their past misinformation slip-ups and the dangers of persuasive deepfakes, social media companies are moving quickly ahead of the 2020 election.