Twitter said tuesday it will begin tagging Twitter messages that contain synthetic or deceptive media content. In the run-up to the 2020 U.S. presidential election, Twitter is bracing for a flood of fake news. Twitter also said it would remove any media content that could be harmful and deliberately misleading, including content that could pose a personal security threat, social disorder, voter suppression or privacy risk.
Social media companies have been under pressure for some time to deal with the emerging threat of Deepfake video. These videos use artificial intelligence technology to create very realistic, but really fabricated content. The content of this type of video is usually what someone said and did, but it’s completely out of line with the real situation.
Alphabet’s YouTube said earlier this week it would remove any content that had been modified and could pose a “serious risk of serious harm.”
Facebook said last month that it would remove Deepfake videos and other deliberately tampered videos from its site, but would retain satirical content and videos in editing that “just omit or change word order.”
Under the new policy, Twitter will label any photos and videos that are “obviously and deceptively edited or fabricated.” However, Twitter does not distinguish between techniques used to modify media content. Yoel Roth, head of honesty at Twitter, said: “In line with this policy, our focus is on results, not how the process works. “
He did not say what resources Twitter would invest in discovering the tampered media content, but said it would consider users’ reports and work with “third-party experts” to identify the content.