Donald Trump's Presidential Campaign Released Deepfakes Attacking Ron DeSantis

Former President Donald Trump released two AI-generated videos using deepfaked voices to mock Florida Governor Ron DeSantis. The first video, posted on platforms like Rumble and Instagram, depicted a chaotic and offensive fake Twitter Spaces event featuring deepfaked voices of Elon Musk, George Soros, Klaus Schwab, Dick Cheney, Adolf Hitler, and a generated voice of Satan. The second video showed a rocket with "Ron 2024" written beside it falling and exploding before liftoff.

Risk classification

• Primary risk domain: 4 Malicious actors
• Primary risk subdomain: 4.1 Disinformation, surveillance, and influence at scale

The campaign used generative AI to create and distribute deepfakes of public figures at scale to influence political processes and public opinion during a presidential primary election.

Additional risk subdomains

• 4.3 Fraud, scams, and targeted manipulation: The campaign used AI voice cloning to impersonate and mock specific individuals, including Ron DeSantis and Anderson Cooper, for political advantage.

Causal factors

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

The incident was caused by the intentional decision of humans within the Trump campaign to generate and distribute deepfaked media to influence public opinion.

EU AI Act risk tier

• Risk tier: 3 Limited Risk

Risk Level 3: Limited Risk. The incident involves the creation and dissemination of AI-generated content such as deepfakes, which are subject to transparency obligations under the EU AI Act to ensure users are aware the content is synthetic.

AI system and alleged parties

• AI system: Midjourney, unspecified (Midjourney)
• AI purpose: Voice Generation; Image Generation
• Behaviour type: Tool
• Alleged developer: Unknown deepfake creators
• Alleged deployer: Donald Trump presidential campaign
• Alleged harmed parties: Ron DeSantis, Klaus Schwab, George Soros, Elon Musk, Dick Cheney

Harm severity

Severity ratings are based on this scale.

Highest direct severity in any category: Minor. 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 Minor, indirect Negligible
• Differential treatment: direct Negligible, indirect Negligible
• Civil rights: direct Negligible, indirect Negligible
• Democracy: direct Minor, indirect Minor
• Privacy: direct Negligible, indirect Negligible
• Psychological: direct Negligible, indirect Negligible
• Epistemic: direct Minor, indirect Minor
• Child sexual exploitation and abuse: direct Negligible, indirect Negligible

Malicious content

Reported: The report explicitly describes the creation and spread of malicious synthetic content, specifically deepfaked voices of public figures used to mock and denigrate them.

Directly caused: The Trump campaign created and distributed a deepfaked video featuring cloned voices of political figures and an AI-generated video of an exploding rocket across platforms like Rumble and Instagram.

Democracy

Reported: The report explicitly describes the erosion of democratic norms, noting that these clips mark the start of a highly combative presidential primary election fueled by generative AI.

Directly caused: The campaign used deepfakes of political opponents and public figures to influence public perception and mock a rival candidate during an active election cycle.

Indirectly caused: The normalization of deepfakes in political campaigns could erode public trust in authentic political discourse and democratic processes.

Inferred additional harm: Continued unlabelled use of deepfakes in elections is highly likely to further degrade trust in democratic institutions and authentic campaign communications.

Epistemic

Reported: The report explicitly describes epistemic harm, noting that the amplification of AI-generated political images and deepfakes on social media can lead to meaningful public confusion.

Directly caused: The campaign released deepfaked audio of public figures, contributing to a confusing information environment where authentic and synthetic media are blurred.

Indirectly caused: The report notes a separate incident where an AI-generated image of an explosion at the Pentagon led to meaningful confusion on social media.

Inferred additional harm: The proliferation of highly realistic deepfakes makes it increasingly difficult for the public to distinguish between real and fabricated media, degrading the shared information ecosystem.

People affected

• Occurrences reported: 1
• People reportedly harmed: 6
• People reportedly exposed: 1000000

Potential causes

Management

• Prioritizing Engagement Over Truth: Campaign management prioritized viral engagement over ethical communication.
• Lack of AI Ethics Policies: Campaigns lacked internal guidelines or oversight for using generative AI.

Technology

• Advanced Voice Cloning Models: Allowed highly realistic replication of public figures' voices without consent.
• Generative Video Tools: Enabled creation of fake campaign videos like the exploding rocket.
• Lack of Platform Detection: Social media lacked tools to flag or block synthetic media automatically.

Data Inputs

• Abundant Public Audio Data: Abundant recordings of public figures enabled training of accurate voice models.
• Unlabeled Synthetic Content: Lack of metadata or watermarks on generated assets facilitated dissemination.

Human Factors

• Intentional Deception by Campaign: Campaign staff deliberately used AI to mock opponents and influence voters.
• Public Susceptibility to Fakes: Audience struggles to distinguish real audio from synthetic political parody.

Process and Methods

• Inadequate Content Moderation: Platforms lacked robust moderation processes to handle political deepfakes.
• Absence of Verification Checks: No pre-upload verification processes for synthetic political advertisements.

Regulatory Environment

• Lack of Deepfake Regulations: Absence of clear laws banning or restricting political deepfakes in campaigns.
• Weak Platform Accountability: No regulatory penalties for hosting unlabeled AI-generated political content.

Information quality

• Classification confidence: High
• Reason for confidence: The report provides clear, detailed accounts of the specific deepfaked videos and images released by the Trump campaign, identifying the public figures targeted and the platforms used. While the exact technical models for the voice cloning are unspecified, the nature of the AI involvement (deepfakes and image generation) is unambiguous.
• Ambiguities identified: The specific AI voice cloning models used by the campaign are not identified.
• Alternative interpretations: The content could be viewed purely as protected political parody rather than a malicious disinformation campaign, though its potential to mislead still poses epistemic risks.

The Donald Trump presidential campaign utilized generative AI voice-cloning tools to release deepfaked videos mocking political rivals during the 2024 primary. While intended as parody, the incident highlights the evolving threat of synthetic media in domestic political campaigns, contributing to epistemic harm and the erosion of democratic norms, though posing minimal direct national security risk.

• Overall national security impact: Minor
• Scope: Single nation
• Primary target: United States
• Alleged perpetrator: Donald Trump presidential campaign

Threat characteristics

• Imminence: Long-term. Represents an ongoing tactical trend in political campaign communication rather than an active national security crisis.
• Autonomy: Human-controlled. AI voice cloning and image generation tools were directly operated and deployed by campaign staff with no autonomous decision-making.
• Novelty: Evolved capability. Represents a significant escalation in the quality, accessibility, and deployment scale of voice-cloning deepfakes within mainstream political campaigns.

Impact by dimension

• Physical security: Negligible. No physical infrastructure, kinetic systems, or human safety elements were targeted or affected by these political deepfakes.
• Information security: Minor. Domestic political campaign used parody voice-cloning and synthetic media, contributing to domestic information manipulation without compromising state intelligence capabilities.
• Sovereignty: Minor. Domestic electoral campaign discourse was targeted with synthetic media, representing minor political manipulation but no direct disruption of core government functions.
• Economic security: Negligible. The incident had no impact on financial systems, strategic technologies, or national economic security.
• Societal stability: Minor. The proliferation of unlabelled political deepfakes contributes to epistemic harm and public confusion, eroding trust in democratic norms and the shared information ecosystem.