The German government has just drafted a bill that would make it a criminal offence to create or share AI-generated content, which could be punishable by up to two years in prison if the content “appears to describe a real event involving another person” and “is likely to cause significant damage to that person’s reputation”.

There is no requirement that the content should be pornographic, violent or offensive in the traditional legal sense. The government decides what is considered “defamatory” and a person who shares a meme can be sentenced to imprisonment.

Federal Justice Minister Stefanie Hubig (SPD) pushed the bill forward after actress Collien Fernandes made allegations against her ex-husband Christian Ulmen concerning deepfake pornography made without consent.

The bill, which Hubig dug out of the ministry’s archives, goes so much further than protecting victims of deep-fake sexual videos that the link seems like a mere pretext.

The draft law proposes three new or extended criminal provisions. The revised Article 184k of the Penal Code would apply to pornographic deepfake videos made without consent, which would be punishable by up to two years’ imprisonment, while the new Article 201b would criminalise non-pornographic deepfake videos that “may cause significant damage to reputation”. New Article 202e would target unauthorised digital surveillance and spying using stalkerware.

It is legitimate for law enforcement to punish people who install secret tracking tools on another person’s phone without their knowledge. However, that provision covers everything else in this bill.

Because the real problem is §201b, a provision that has nothing to do with pornography, but deals exclusively with political expression. As criminal lawyer Udo Vetter wrote in his law blog, the proposed §201b “punishes with a prison sentence of up to two years anyone who makes available to a third person content created or edited by a computer program that gives the impression of reflecting an actual event concerning another person and is likely to cause significant damage to that person’s reputation.”

This wording is broad enough to cover a satirical meme about a politician, a parody video or a manipulated image used for political commentary.

Vetter calls the bill “a fresh polished Trojan horse”. His central point is damning. Proposed §201b “does not require the display of naked persons or body parts or the placing of persons in sexually degrading positions. The section does not require degradation, glorification of violence, or incitement to hatred. It simply targets ‘substantial damage to reputation’ caused by a meme or video”, a standard that goes well beyond deepfake pornography and into the heart of political discourse.

However, there is an exception clause in the draft law for satire and political art. In theory, it looks like a protective measure. According to Vetter, the exception is “structurally useless” for the central genre of political satire, since the whole point of political deepfake is precisely that it imitates reality.

The satirical value comes from the similarity with real events. A court decision that a particular meme meets the criteria for protected satire may not be issued until years later. A police raid on the home of the meme’s creator takes place at 6am.

This is a frightening effect, and Vetter mentions it explicitly. Not a judgment. It’s a process. “The confiscation of all computers and equipment, the legal costs and months of stigma from an ongoing criminal investigation are punishment enough in themselves,” he writes.

“Legitimate, sharp criticism is dampened by fear of repression alone, before any court has yet made a decision.” People stop sharing political memes not because they have been found guilty, but because they are afraid to take the risk.

Germany already has direct experience of just such abuse. The 2021 law on insulting politicians (§188 StGB) led in 2024 to a house search of a man who had shared an ironic online photo collage calling the then Economics Minister Robert Habeck a “Schwachkopf” (idiot) in the font of the cosmetics brand Schwarzkopf. The law, which was intended to protect councillors from harassment, was used to send the police to someone’s home over a joke about a federal minister. The proposed deepfake law would give prosecutors even broader powers.

The German Bar Association (DAV) has already said that the bill “goes too far”. Ali B. Norouzi, chairman of the DAV’s criminal law committee, warns of “urgent legislative action” due to “public anger fuelled by the media”.

DAV board member Niko Härting previously told LTO that it was “doubtful” whether there were any gaps in criminal liability issues related to deepfake technology that would justify new legislation, pointing out that personality rights are already protected under existing defamation laws.