Microsoft takes the field against disinformation by targeting deepfakes, i.e. videos created with a technique that uses artificial intelligence to superimpose a different face on the face of a person shot in a movie, creating an exchange of identity.
The Redmond company has unveiled a tool called Video Authenticator, which scans footage to find manipulated ones.
Video Authenticator can be used to analyze photos and videos, and it provides a given percentage of the likelihood that those images are fake, Microsoft explains in a statement.
In the case of movies, the percentage is relative to each single frame.
The new tool, which comes a few weeks after the US presidential election, aims to tackle the misinformation that sees deepfakes as "one of the biggest problems", because "they can make it seem like a person is saying things they never said. or that it's in a place where it really wasn't, "Microsoft notes.
The fact that deepfakes are generated by artificial intelligence makes it "inevitable" that these manipulations will beat the technologies aimed at detecting them.
"However - the company emphasizes - in the short term, for example for the upcoming US elections, advanced detection technologies can be a useful tool to help savvy users identify deepfakes".