This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
A viral deepfake video circulated on social media platforms, including TikTok, X, and Facebook, falsely portraying U.S. President-elect Donald Trump calling for the release of Nigerian separatist leader Nnamdi Kanu. The video used old footage of Trump paired with AI-generated audio and contained factual errors, such as a reference to 'November 31st.' Fact-checkers confirmed the video was inauthentic, highlighting the risks of AI-generated misinformation in political contexts.
A reportedly AI-manipulated video circulated on social media that falsely portrayed U.S. President-elect Donald Trump calling for the release of Nigerian separatist leader Nnamdi Kanu and threatening sanctions against Nigeria. The video allegedly used old footage paired with AI-generated audio mimicking Trump's voice. Fact-checkers identified the video as inauthentic, noting mismatched lip-sync, fabricated quotes, and the fictitious date of "November 31st."
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
Using AI systems to conduct large-scale disinformation campaigns, malicious surveillance, or targeted and sophisticated automated censorship and propaganda, with the aim of manipulating political processes, public opinion, and behavior.
Also classified as
The deepfake generated false information attributing a fabricated political stance and threat of sanctions to a US President-elect.
The deepfake was created to conduct a targeted political disinformation campaign aimed at manipulating public opinion and political processes regarding the detention of a Nigerian separatist leader.
Human
Due to a decision or action made by humans
Intentional
Due to an expected outcome from pursuing a goal
Post-deployment
Occurring after the AI model has been trained and deployed
The incident was caused by a human actor intentionally using a deployed AI voice generation tool to create a political deepfake.
Added categories peak at 2: Minor (Epistemic Harm, Direct)
A viral deepfake video falsely depicted US President-elect Donald Trump threatening the Nigerian government with sanctions unless they released a separatist leader. While targeting sensitive bilateral relations and exploiting political tensions, the low-quality video was quickly debunked, resulting in minor national security impact.
The AI system was used as a tool by a human creator to generate specific synthetic audio mimicking Donald Trump's voice based on direct inputs, without exhibiting autonomous planning or agency.
Limited Risk: The report describes an AI-generated deepfake video, which falls under the transparency obligations of the EU AI Act for AI-generated content.
The AI system was used to generate a deepfake video with synthetic audio mimicking Donald Trump's voice.
The fact-checking report is clear, detailed, and provides direct evidence of the deepfake's existence, its platform reach, and the specific false claims made. The verification process is well-documented, leaving little ambiguity about the nature of the incident.
None. The evidence clearly points to an intentional political deepfake.
Technology
AI Voice Cloning
AI tools allowed creation of highly realistic voice mimicking Trump.
Video Splicing Tools
Editing tools enabled splicing unrelated footage with synthetic audio.
Data Inputs
Unrelated Video Footage
Visuals were clipped from an unrelated event of Trump speaking.
Synthesized Audio Input
AI-generated audio was used to mimic Trump's voice speaking on IPOB.
Human Factors
Political Motivation
Supporters sought international backing, motivating them to share.
Lack of Critical Evaluation
Users believed the video despite obvious errors like November 31st.
Process and Methods
Viral Social Media Sharing
Algorithms and sharing features allowed rapid spread across platforms.
AI system involved
An AI system was involved, but the reports do not identify it.
Identified by automated extraction; can differ from the alleged developers above.
