This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
In December 2025, a deepfake image of U.S. President Donald Trump using a walker circulated widely on social media platforms, including X, Facebook, and Threads. The image was shared by a political strategist to falsely suggest the President was frail following the signing of an AI executive order. Fact-checkers and AI detection tools, including Google's Gemini and Hive, confirmed the image was AI-generated and contained a SynthID watermark, highlighting the potential for generative AI to be used in spreading political disinformation.
A fabricated image purporting to show U.S. President Donald Trump using a walker reportedly circulated widely on social media in mid-December 2025. Fact-checkers determined the image was AI-generated, citing visual anomalies and detection of Google's SynthID watermark. The image reportedly spread across multiple platforms and was shared by a political strategist, prompting public confusion and debunking by news outlets and AI detection tools.
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 AI image generator lacked safeguards to prevent the generation of deceptive images of public figures.
The incident involved the creation and viral dissemination of an AI-generated deepfake of President Trump to influence public perception of his health and political actions.
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 initiated by a human decision to generate and intentionally share a deceptive deepfake of a political figure on social media.
Added categories peak at 4: Severe (Epistemic Harm, Indirect)
A political strategist circulated an AI-generated deepfake of President Trump using a walker to suggest physical frailty. While the image went viral, it was quickly debunked by AI detection tools and fact-checkers. The incident highlights ongoing concerns regarding safeguards in commercial image generators but presents minor national security risk.
The report describes the AI system as an image generator used to perform a specific function based on direct user prompts, operating as a tool without independent planning or autonomous decision-making.
Limited Risk: The system is an AI-generated content generator used to create deepfakes, which under the EU AI Act poses a moderate risk and requires specific transparency obligations such as watermarking.
The AI system was used to generate a highly realistic, deceptive image of a public figure (Donald Trump) using a walker.
The reports provide clear, consistent details about the origin of the image, the platforms where it circulated, the specific AI model used (Gemini Nano Banana Pro), and the verification of the SynthID watermark. There are no major conflicting accounts, and the facts are well-documented by multiple fact-checking organizations.
The incident could be interpreted as a harmless political joke or satire rather than a malicious disinformation campaign, though its scale and the fact that many took it seriously support the disinformation framing.
Technology
Lack of Public Figure Safeguards
Nano Banana Pro lacked safeguards preventing generation of public figures.
Absence of Visible Watermarks
No visible watermark on the AI image made immediate identification difficult.
Highly Realistic AI Generation
Advanced generation tools produce highly convincing fake images of people.
Human Factors
Sharing Unlabeled Satirical Content
Strategist posted the AI image as 'breaking' without labeling it as satire.
User Confirmation Bias
Liberal social media users shared the image to highlight Trump's frailty.
Lack of Public Verification
Users failed to inspect the image for AI artifacts before reposting.
Process and Methods
Slow Social Media Moderation
Platforms allowed the fake image to spread to millions before debunking.
Complex Watermark Detection
SynthID digital watermark requires external AI tools to be detected.
Management
Inadequate Product Safeguards
Google released Nano Banana Pro without restricting public figure generation.
Regulatory Environment
Weak AI Regulation Standards
Executive order focused on limiting state rules rather than safety standards.
Lack of Generative AI Mandates
No regulatory mandate requiring visible labeling on AI-generated media.
AI system involved
Identified by automated extraction; can differ from the alleged developers above.
