Democracy & information IWIB4AI session Needs review
Generative AI can amplify election misinformation, challenging democratic trust while demanding safeguards that protect free expression.
Draft – legal and case law references not yet verified
This use case examines how AI-generated video, synthetic audio and text can amplify misinformation and undermine electoral integrity. Convincing fabricated content can impersonate political figures, mislead voters and weaken trust in authentic information. Responses include content transparency, platform risk mitigation and international cooperation through initiatives such as the non-binding UN Global Digital Compact.
Generative AI makes misleading political content easier to produce and distribute at scale, while voters may struggle to distinguish authentic material from fabrications. The central challenge is protecting informed electoral participation without suppressing legitimate political expression, satire or debate.
Fabricated statements and misleading election information can distort public debate and interfere with informed participation. Safeguards should protect the electoral process rather than advantage particular political viewpoints.
Deceptive synthetic content can manipulate voters by misrepresenting who said something or what occurred. Voters need reliable context to form political judgments independently.
Clear disclosure of synthetic content helps voters assess its origin and credibility. Labels should not imply that all AI-generated content is false or that unlabelled content is necessarily authentic.
AI providers, content publishers and platforms have distinct responsibilities for preventing and addressing electoral deception. Moderation and mitigation measures should be explainable and open to challenge to reduce arbitrary restrictions on political speech.
| Framework | Provision | Relevance |
|---|---|---|
| EU AI Act | Art. 50(2), (4) | These provisions establish machine-readable marking obligations for providers of systems generating synthetic content and disclosure duties for deployers producing deepfakes or public-interest text, subject to specified exceptions and qualifications. |
| DSA | Arts. 34–35 | Very large online platforms and search engines must assess and mitigate systemic risks, including actual or foreseeable negative effects on civic discourse and electoral processes. |
| GDPR | Arts. 5, 6 and 9 | Where electoral AI activities process personal data, they require lawful, fair and transparent processing, with additional restrictions for data revealing political opinions. |
| EU Charter of Fundamental Rights | Arts. 11 and 52(1) | Within the Charter’s scope, measures restricting electoral content must respect freedom of expression and information and satisfy the requirements governing limitations on fundamental rights. |
Annulment of the 2024 Romanian presidential election Needs review
The annulment illustrates how coordinated online manipulation can threaten electoral legitimacy, without establishing that generative AI caused the interference.
Use Case #5 – GenAI's Impact on Election Integrity & Democracy
