Significant increase in manipulated AI videos: This is how YouTube wants to combat it

Last year, 35,000 videos that spread disinformation in this way were blocked, according to a background discussion with journalists. For the EU election campaign, YouTube wants to focus on labeling AI content.

• Also read: Believe your eyes and ears – don’t [OÖNplus]

Every video that contains AI content should be marked with an AI label, similar to advertising (“altered or synthetic content”). If manipulated content is uploaded without a label, YouTube will add it later or block the video. There are no exceptions; content from politicians must also be labeled. This is necessary because foreign actors, specifically Russia, want to influence the election campaign.

Deepfake to deceive voters

AI and deepfake representations are being used by these actors to deceive voters, it said. Users are particularly often deceived by old video recordings being passed off as new ones. This trend is not new. YouTube representatives said something similar had already been observed in the Ukraine war and the Middle East conflict. A high five-digit amount of content has already been deleted here.

• You might also be interested in: “Artificial intelligence also needs people who can handle it” [OÖNplus]

In addition to curbing disinformation, YouTube also wants to increase political participation in the EU elections. Under the motto “Supporting the European Parliamentary Elections” the aim is to provide users with targeted content such as information about the candidates, the electoral system and the general conditions.

Users should also be reminded to exercise their right to vote through advertisements. Furthermore, the plan is to put “trustworthy sources” in the foreground. They are more likely to appear in searches than other sources. When asked, YouTube representatives emphasized that this meant established media such as daily newspapers and TV stations.