EU Asks Facebook, TikTok to Identify and Label AI Deepfakes
Facts
- The European Commission (EC) has, under the Digital Services Act (DSA), called for 'Very Large Online Platforms' and 'Very Large Search Engines' to 'mitigate systemic risks,' such as artificial intelligence (AI) deepfakes, ahead of upcoming European Union (EU) elections in June.1
- Affected companies include Facebook, X, YouTube, and TikTok. The EC's guidelines require political ads and deepfakes to be labeled on the platforms.2
- The EC has released guidelines concerning how mitigation and compliance with the DSA may occur. While the guidelines are not legally binding, the EC has stated that companies must prove that other measures are 'equally effective,' and that stress tests to platform provisions will be carried out in April.3
- The guidelines describe the large platforms and search engines as services that 'bear risks of actual or foreseeable negative effects on electoral processes,' while recommending independent analyses of internal processes.4
- The DSA's obligations for companies to apply protections ahead of the June elections concerns all platforms that contain over 45M active users, while EC Vice Pres. Vera Jourova has also encouraged all political parties to refrain from AI and deepfake use.5
Sources: 1European Commission, 2DW, 3POLITICO, 4EEC.Europa.EU and 5Bloomberg.
Narratives
- Pro-establishment narrative, as provided by POLITICO. Deepfakes and other forms of disinformation threaten the integrity of the upcoming European elections. The June contests will likely be the target of foreign interference and hybrid attacks so protective measures must be put in place. It is imperative that social media platforms do their best to regulate this content on their sites and mitigate the risks to the election.
- Establishment-critical narrative, as provided by Reason. The DSA is an enormous example of government overreach. While it's admirable that the EU is attempting to set rules that could provide some order to the internet, these powers can be used to infringe on the free speech rights of Europeans and others and unfairly censor the content on social media platforms.