The use of artificial intelligence to generate deepfakes has surged, with detected cases rising from half a million to over 8 million between 2023 and 2025, according to the cybersecurity firm DeepStrike. These tools now allow individuals to create fake images, videos, and audio without technical expertise. The proliferation of this technology has prompted urgent calls for legislative action in Switzerland to combat digital abuse, sexual exploitation, and the destabilization of democratic processes.
The Rising Scale of Digital Manipulation
Deepfakes are increasingly used for malicious purposes, ranging from financial fraud to targeted harassment. Scammers utilize manipulated videos of celebrities to promote fraudulent investment schemes, while the creation of non-consensual sexual imagery—often referred to as “nudification”—has become a widespread concern. The Swiss platform clickandstop.ch, which tracks cyberharassment and sexual abuse involving children, reported a 63% increase in incidents in 2025. Data from the platform, supported by Child Protection Switzerland, indicates that victims are increasingly reporting the circulation of AI-generated nude photos.
Did You Know? The number of detected deepfake cases worldwide surged from 500,000 to more than 8 million in just two years, according to figures from the American cybersecurity company DeepStrike.
Legislative Efforts and Political Debate
Swiss authorities remain divided on the best approach to regulate this technology. In June, the government and the National Council approved a motion from Green Party deputy Raphaël Mahaim focused on sexual deepfakes. The proposal suggests requiring AI platforms to perform risk assessments and implement protective measures, backed by a potential oversight body capable of issuing sanctions.
However, some officials advocate for utilizing existing frameworks rather than creating new, specific laws. Albert Rösti, head of the Federal Department of the Environment, Transport, Energy and Communications (DETEC), has previously stated that current Swiss penal and civil codes offer sufficient protection. He also noted Switzerland’s participation in the Council of Europe’s AI Convention, which aims to protect human rights and the rule of law.
Expert Insight: There is a significant tension between the desire for rapid legislative protection and the need for legal clarity. David Vasella, founder of the platform datenrecht.ch and associate member of the law firm Walder Wyss, suggests that layering new laws specifically for deepfakes could complicate the existing legal landscape, proposing instead that the government empower current oversight bodies, such as the Federal Data Protection and Information Commissioner, to handle these emerging threats.
Platform Accountability and Technical Solutions
Multiple legislative paths are currently under consideration. A group of cross-party parliamentarians is backing an “internet initiative” that could lead to a public vote on protecting fundamental rights in the digital space. Separately, the Swiss government has drafted legislation—first discussed in 2023 and submitted in October—that would require large social media platforms and search engines to formally process complaints regarding defamation, discrimination, and hate speech. Under this proposal, companies would need to maintain legal representation in Switzerland and publicly justify decisions to remove content or block accounts.
Beyond legislation, researchers are developing technical defenses. At the Swiss Federal Institute of Technology Zurich (EPFZ), scientists are working on hardware-based authentication. By integrating encrypted digital signatures directly into camera sensors, the team aims to verify the authenticity of content at the point of creation. “It is crucial to generate the signatures directly on the chip of the data sensor,” says Fernando Cardes, a member of the EPFZ research team, noting that this method makes falsification significantly more difficult than software-based approaches.
Frequently Asked Questions
What is the primary concern regarding deepfakes in the political sphere?
Deepfakes are increasingly used to manipulate public perception by enhancing a candidate’s image or targeting opponents. Furthermore, there is growing alarm regarding the use of these tools by foreign states, such as Russia, to destabilize Western democracies.
What does the government’s proposed legislation require of social media platforms?
The draft law would force major platforms to provide a formal process for handling complaints about defamation, insults, discrimination, and incitement to hate. Platforms would also be required to establish legal representation in Switzerland and justify their content moderation decisions publicly.
How does the EPFZ’s detection technology differ from current methods?
While many existing systems use software-based signatures, the EPFZ researchers have developed sensor-based technology. By generating a digital watermark directly on the camera chip at the moment of capture, they believe they can close the security gaps that allow hackers to manipulate content after it has been created.
How should society balance the rapid evolution of AI tools with the protection of individual privacy and democratic integrity?
Worth a look
