The Scoop
German government officials are exploring ways to push social media platforms to add authentication tools that disclose whether an image is real or manipulated.
The discussions are still in the very early stages, Benjamin Brake, head of digital policy in Germany’s Federal Ministry for Digital and Transport, told Semafor. Ministry officials haven’t yet approached social media platforms about the idea, but have met with a company that provides image authentication tools, he said.
It’s likely the issue may ultimately need to be addressed on a European Union level, rather than just Germany, he added.
“We would like to try to find a way that people actually can identify if something is real,” Brake said.
It’s too early to say how such a tool would work in practice, and whether it would be legally required, but Brake said it would likely be a kind of metadata tag that accompanies photos and would be different from watermarking, a practice in which a signal is embedded into an image that identifies it as generated by artificial intelligence.
“We’re trying to talk to huge online platforms, [about] why they are not using it, if there are any obstacles, and in that way, try to push them a bit into… at least working with the images that are proliferating over their platform,” he said.
Know More
Governments globally have sounded the alarm about fake or manipulated images circulating on social media, especially during elections. Germany is holding three closely watched state votes in September — two of them on Sunday — and a federal election is scheduled for September 2025.
EU regulators are trying to crack down on misleading or harmful content through the Digital Services Act, which requires so-called very large online platforms to have systems in place to handle misinformation, hate speech, and terrorist activity. But mandatory image authorization tools aren’t part of the EU’s current rules.
Fake AI-generated content has been a particular challenge. Russian influence campaigns spread deepfakes ahead of European Parliament elections in June, and more recently, unknown actors used the identities of European influencers to push pro-Donald Trump messaging in the US, CNN reported Wednesday.
Some platforms, like TikTok, and Meta’s Instagram and Facebook, require users to add a label to AI-generated content, and have recently deployed tools aimed at identifying when AI is used.