Parody AI Images of Donald Trump Being Arrested Reposted as Misinformation

AI-generated photorealistic images depicting Donald Trump being detained by the police which were originally posted on Twitter as parody were unintentionally shared across social media platforms as factual news, lacking the intended context.

Risk classification

• Primary risk domain: 3 Misinformation
• Primary risk subdomain: 3.1 False or misleading information

The AI system generated highly realistic but entirely fabricated images of a public figure's arrest, which were subsequently spread online and misled some users into believing a false event had occurred.

Additional risk subdomains

• 4.1 Disinformation, surveillance, and influence at scale: The synthetic images were shared widely across multiple platforms, demonstrating the potential for AI-generated media to be used for large-scale political influence and public opinion manipulation.

Causal factors

• Entity: Human
• Intent: Intentional
• Timing: Post-deployment

The incident was initiated by a human user intentionally prompting the AI to generate the images, and further propelled by other humans intentionally sharing them across social media platforms.

EU AI Act risk tier

• Risk tier: 3 Limited Risk

Limited Risk: The incident involves AI-generated deepfakes, which under the EU AI Act are subject to specific transparency and disclosure obligations to prevent user deception.

AI system and alleged parties

• AI system: Midjourney 5 (Midjourney)
• AI purpose: Deepfake Image Generation; Visual Art Generation
• Behaviour type: Tool
• Alleged developer: Midjourney
• Alleged deployer: Eliot Higgins
• Alleged harmed parties: Twitter Users, social media users

Harm severity

Severity ratings are based on this scale.

Highest direct severity in any category: Substantial. Severity is scored from Negligible to Catastrophic in each harm category, for harm the reports describe as caused directly or indirectly by the AI system.

• Physical: direct Negligible, indirect Negligible
• Infrastructure: direct Negligible, indirect Negligible
• Property: direct Negligible, indirect Negligible
• Financial: direct Negligible, indirect Negligible
• Environmental: direct Negligible, indirect Negligible
• Malicious content: direct Negligible, indirect Negligible
• Differential treatment: direct Negligible, indirect Negligible
• Civil rights: direct Negligible, indirect Negligible
• Democracy: direct Negligible, indirect Minor
• Privacy: direct Negligible, indirect Negligible
• Psychological: direct Negligible, indirect Negligible
• Epistemic: direct Substantial, indirect Substantial
• Child sexual exploitation and abuse: direct Negligible, indirect Negligible

Democracy

Reported: The reports describe concerns about the erosion of democratic discourse and public trust in visual evidence.

Indirectly caused: The viral spread of the fake arrest images of a former president during an active investigation created confusion in volatile news environments and threatened to undermine trust in real news and visual evidence.

Inferred additional harm: Widespread public cynicism and further erosion of trust in democratic institutions and media verification processes.

Epistemic

Reported: The reports explicitly describe epistemic harm through the creation and spread of highly realistic fake images depicting a major political event that did not occur.

Directly caused: Midjourney v5 was used to generate photorealistic images of Donald Trump being arrested, which were viewed over 5 million times on Twitter and spread to other platforms like Facebook and Instagram.

Indirectly caused: Some social media users shared the images as genuine, leading to public confusion and the need for extensive fact-checking by platforms and news agencies.

Inferred additional harm: Increased difficulty for the general public to distinguish real from synthetic media, leading to a general degradation of the shared information ecosystem.

People affected

• Occurrences reported: 1
• People reportedly exposed: 5000000

Potential causes

Management

• Tech Industry Layoffs: Massive staff cuts reduced trust and safety teams at major platforms.
• Inadequate Safety Safeguards: Midjourney released advanced v5 without robust prompt filters or warnings.
• Opaque Enforcement Actions: Midjourney banned the creator without explaining which rules were violated.

Technology

• Rapid Generative AI Sophistication: Midjourney v5 released with advanced photorealism mimicking news style.
• Imperfect Rendering of Details: AI still struggles with hands, text, and extra limbs, causing some odd details.
• Ease of Generating Synthetic Media: Low cost and simple text prompts allow rapid creation of realistic images.

Data Inputs

• Scraped Celebrity Training Data: Vast public training data makes famous figures render highly realistically.
• Lack of Image Provenance Data: Generated images lack embedded metadata or watermarks to verify origin.

Human Factors

• User Sharing Without Context: Users reposted parody fakes as real news to validate political biases.
• Lack of Critical Thinking: Public struggled to spot visual errors like extra limbs or garbled text.
• Creator Experimentation: Users generated provocative political scenarios to test tool capabilities.

Process and Methods

• Inconsistent Platform Policies: Social media sites had varied and weak rules on synthetic media.
• Inadequate Detection Tools: Platforms lacked automated systems to detect and label AI fakes.
• Lack of Contextual Labeling: Fakes quickly lost their original labels when shared across platforms.

Regulatory Environment

• Absence of AI Regulations: No government policies or standards exist to govern synthetic media creation.
• Lack of Developer Liability: Developers face no legal consequences for foreseeable harms of their tools.

Information quality

• Classification confidence: High
• Reason for confidence: The reports provide consistent, detailed accounts of the creation, spread, and impact of the Trump arrest deepfakes, including statements from the creator, AI experts, and platform responses.
• Ambiguities identified: None of significance; the facts of the event are clear and verified.
• Alternative interpretations: The incident could be interpreted as a benign artistic parody rather than a safety risk, but the documented public confusion and platform policy violations support the misinformation classification.

The viral spread of AI-generated images depicting the fictional arrest of Donald Trump highlighted the evolving threat of photorealistic deepfakes. While the incident was initiated as a parody by a journalist and did not cause physical or systemic harm, it demonstrated how easily synthetic media can pollute the information ecosystem and confuse the public, representing a minor national security concern.

• Overall national security impact: Minor
• Scope: Single nation
• Primary target: United States
• Alleged perpetrator: Eliot Higgins

Threat characteristics

• Imminence: Long-term. The immediate incident was resolved through fact-checking, but the capability represents an ongoing strategic challenge.
• Autonomy: Human-controlled. The AI tool was directly prompted and controlled by a human creator without autonomous execution.
• Novelty: Evolved capability. Represents a significant advancement in the photorealism and accessibility of generative image tools compared to earlier models.

Impact by dimension

• Physical security: Negligible. No physical threats, kinetic actions, or critical infrastructure impacts were reported.
• Information security: Minor. Demonstrated the potential for high-quality synthetic media to spread rapidly and cause public confusion, though it was not a coordinated foreign intelligence operation.
• Sovereignty: Negligible. Did not disrupt core government operations, electoral infrastructure, or state sovereignty.
• Economic security: Negligible. No financial system impacts or strategic technology theft occurred.
• Societal stability: Minor. Generated public confusion regarding the arrest of a political figure, highlighting risks to democratic discourse, but did not spark civil unrest.