A political majority in the Danish Parliament has reached a new agreement on the use of deepfakes in political campaigns and election campaigns. The voluntary agreement, which all parties in the Parliament except for the Danish People’s Party and Liberal Alliance have signed, means that parties must refrain from using deepfakes to depict or copy politicians or political candidates. Deepfake technology uses artificial intelligence to manipulate images or videos, which can create very realistic, but false, representations. In a press release, the Speaker of the Parliament, Søren Gade, expresses that the use of artificial intelligence in political communication must never come at the expense of credibility and transparency. He is pleased that the parties now have a common understanding of the rules for deepfake use. The code of conduct, which came into effect immediately, is voluntary, but the Danish People’s Party and Liberal Alliance have chosen not to join. These two parties have previously used deepfakes, for example, the Danish People’s Party manipulated a video with Prime Minister Mette Frederiksen, where she, among other things, says that the government will abolish more holidays. The video stated that it was created using artificial intelligence. Mette Frederiksen has criticized the use of deepfakes and called it an “unfortunate slippery slope”. The agreement also entails that any disagreements about deepfake content should be handled internally and in confidence. The new code will be integrated into the existing 2019 agreement on not spreading fake news. After the summer break, the parties will continue to work on a more general code of conduct on responsible use of artificial intelligence in political communication, which will apply to all parties. This initiative represents a step towards more responsible use of technology in political contexts, at a time when digital manipulation can quickly spread misinformation.

Read the danish version here