Twitter has started laying the groundwork for combating the proliferation of "deepfakes" and the hoax videos they produce, which are videos made up with powerful and durable editing tools that look like the real thing but are actually almost entirely made from a fake.
The idea is that once users are targeted with a misleading video, the anti-deepfakes code creates a noise signal that can be automatically detected by tech companies using a custom hashtag on their platforms that links a video creator to the company's verified account.
The company has identified a number of deepfakes accounts that are already working to repackage and retweet promoted disinformation using the "verified hashtag." In another test, bots on Twitter appear to be using a fake account to share content, which the software detects as a false post.
This level of detection, which will take time to resolve, will give tech companies a decent chance of preemptively dealing with a growing number of misinformation-generating videos that could potentially get a large audience on social media.
Deepfakes, which first took off in the wake of the U.S. presidential election, are videos made with elaborate tools that can alter facial expressions, instead of the text and words that we've become used to seeing online. They can also replace a video's footage, which is currently made of stills from videos using videos of different humans and that trigger the same facial movements.
According to Twitter's enforcement blog, the feature will also give the company clues about its opponents' flags that are needed to deploy them with greater success in the future.
"Deepfakes can either be made by bots or edited by humans—in either case, we'll be able to classify these posts as fake," the blog says. "For example, it may indicate that the video is from another site or it may say something suspicious about the creator's account, like an account. This feature will be a good signal of a fake video."
The company will be rolling out the feature slowly, but it will eventually be introduced in every tweet.
The feature could help fighters of Russian interference in the 2016 presidential election and other democratic processes more effectively fight these issues, and hopefully do so with increased efficiency and funding.