Faced with the specter of being banned from social networks, TikTok is trying to adapt its content policy to the requirements of Western countries. The social network has also announced that it will ban the use of deepfakes made using AI.
For several weeks, TikTok has been walking on a really thin wire. Several countries, the United States in the lead, are putting pressure on the social network so that it is sold by Bytedance, the Chinese parent company of the application. It must be said that the links between this company and the Chinese government raise many questions, in particular on the potential use of TikTok to spy around the world.
The social network seeks to reassure by modifying its content management policy: TikTok updated on March 21, 2023 with some important advances, in particular on deepfakes made by AIs.
So far, these deepfakes, which have enjoyed overwhelming popularity in recent months, have managed to avoid the social network’s moderation rules. The only rule that could be found on TikTok regarding deepfakes was limited to a simple line, in which it was explained that deepfakes could “ mislead users by distorting the truth of events [ou] causing significant harm to the subject of the video “.
A simple warning, especially since the application did not seek to remove deepfakes or to sanction users who use them.
More control
But the new rules, updated on March 21, 2023, are more restrictive. Now, TikTok only allows “ synthetic media that presents a public figure in a given context “, but also the artistic and educational deepfakes, in which one could see “ a celebrity performing a trending dance on TikTok “.
But the framing of deepfakes does not stop there. To publish them, they must now include a sticker or a caption, in which it must be indicated that the content is “synthetic”, “false”, “unreal” or “modified”.
Conversely, the use of deepfake is no longer authorized when they contain:
- Realistic scenes that are not denounced or attached clearly in the video.
- The portrait (visual or audio) of a real person, used for political or commercial purposes, or which violates any other policy (hate speech, sexual exploitation or harassment). This includes minors, adult private figures and adult public figures.
- Content edited, edited or combined in such a way that it is likely to mislead a person about real events.
But if Tiktok finally demonstrates a desire to better fight against deepfakes and their propensity to misinform, there is no guarantee that the social network will really sanction, or even banish Internet users who dare to defy “the forbidden”. The moderation system of the app is often indecipherable.
Do you want to know everything about the mobility of tomorrow, from electric cars to pedelecs? Subscribe now to our Watt Else newsletter!