A Purported Deepfake of Senator Elizabeth Warren Circulates Allegedly Saying Republicans Should Not Vote

In February 2023, a purported deepfake of Senator Elizabeth Warren circulated on social media in which allegedly doctored footage of her from an MSNBC interview had her claiming that she believes Republicans should not vote.

Risk classification

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

The incident involves the creation and dissemination of a deepfake video of a US Senator to spread political disinformation and influence public opinion regarding the 2024 election.

Additional risk subdomains

• 3.1 False or misleading information: The AI system was used to generate a highly realistic but entirely fabricated video of a politician making false statements, leading to epistemic confusion among some viewers.

Causal factors

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

The incident was caused by a human actor intentionally using a deployed AI deepfake generation tool to create and distribute misleading political content.

EU AI Act risk tier

• Risk tier: 3 Limited Risk

Limited Risk: The incident involves a deepfake video, which is explicitly classified under Risk Level 3 (Limited Risk) due to transparency obligations requiring users to be informed that the content is AI-generated.

AI system and alleged parties

• AI system: None named
• AI purpose: Deepfake Video Generation; Voice Generation
• Behaviour type: Tool
• Alleged developer: Unknown voice cloning technology developers, Unknown deepfake technology developers
• Alleged deployer: Unnamed deepfake creator, Twitter, TikTok
• Alleged harmed parties: Republicans, MSNBC, General public, Elizabeth Warren, Election integrity, Democrats, Democracy

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 Negligible
• Child sexual exploitation and abuse: direct Negligible, indirect Negligible

Malicious content

Reported: The report explicitly describes the creation and spread of malicious deepfake content designed to mislead the public.

Directly caused: An AI-generated deepfake video of Senator Elizabeth Warren was created and spread on TikTok and Twitter, accumulating over 189,000 views.

Inferred additional harm: It is likely that the video was reshared on other platforms or private messaging apps, exposing more users to the malicious content.

Democracy

Reported: The report explicitly describes the erosion of democratic norms through the spread of political disinformation aimed at polarizing voters.

Directly caused: The deepfake video falsely depicted a US Senator advocating for the disenfranchisement of Republican voters, which directly fuels political polarization and distrust in democratic processes.

Indirectly caused: The video plays into existing narratives that Democrats are attempting to stop Republican candidates through undemocratic means, potentially undermining trust in the 2024 election.

Inferred additional harm: Continued proliferation of such deepfakes could lead to broader public skepticism of authentic political communication and further erode trust in democratic institutions.

Epistemic

Reported: The report explicitly describes epistemic harm through the creation of a false video that successfully duped some viewers into believing it was authentic.

Directly caused: The deepfake video fabricated statements from Senator Warren, misleading some social media users who believed the CGI version was genuinely claiming Republicans should be barred from voting.

Inferred additional harm: The incident contributes to the broader degradation of the information ecosystem, making it harder for citizens to distinguish between real and synthetic political media.

People affected

• Occurrences reported: 1
• People reportedly harmed: 1
• People reportedly exposed: 189800

Potential causes

Management

• Underinvestment in Trust and Safety: Platform management failed to prioritize proactive deepfake detection.

Technology

• Generative AI Deepfake Tools: AI tools allowed realistic synthetic audio and video generation.
• Algorithmic Amplification: Platform algorithms promoted the sensational deepfake, increasing views.

Data Inputs

• Misuse of Authentic Media: Original MSNBC broadcast footage was used as input for the deepfake.

Human Factors

• Malicious Creator Intent: An actor actively sought to spread political disinformation using AI.
• User Confirmation Bias: Some social media users believed the fake because it fit their views.

Process and Methods

• Reactive Labeling Processes: Platforms only labeled the video after it gained significant views.
• Ineffective Upload Screening: Automated screening failed to catch the synthetic content at upload.

Regulatory Environment

• Absence of Deepfake Laws: Lack of regulatory frameworks targeting AI-generated political fakes.

Information quality

• Classification confidence: High
• Reason for confidence: The report clearly details the nature of the deepfake, the platform metrics (views on Twitter and TikTok), the target (Senator Warren), and the political context. There is no ambiguity about the fact that the video was AI-generated and intended to mislead.
• Ambiguities identified: The specific AI tool or software used to generate the deepfake is not identified.
• Alternative interpretations: None. The event is clearly a political deepfake disinformation incident.

An AI-generated deepfake video of US Senator Elizabeth Warren falsely claiming Republicans should be barred from voting circulated on social media. While the video aimed to fuel political polarization and manipulate electoral narratives, its reach was limited, and platform labeling mitigated its overall national security impact, representing a minor information warfare incident.

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

Threat characteristics

• Imminence: Long-term. Represents an ongoing strategic concern regarding the proliferation of political deepfakes rather than an active, imminent crisis.
• Autonomy: Human-controlled. The AI system was used as a tool by a human creator who directed the generation and distribution of the content.
• Novelty: Established threat. Political deepfakes and AI-generated disinformation campaigns are established threat vectors with numerous prior precedents.

Impact by dimension

• Physical security: Negligible. No physical security, kinetic threats, or critical infrastructure damage reported in connection with this incident.
• Information security: Minor. AI-generated deepfake targeted a US Senator to spread political disinformation, but had limited reach under 190,000 views and was quickly labeled by platforms.
• Sovereignty: Minor. The deepfake aimed to influence electoral narratives regarding the 2024 election, but did not disrupt actual government operations or electoral systems.
• Economic security: Negligible. No financial systems, strategic technologies, or economic assets were targeted or affected by this incident.
• Societal stability: Minor. The content sought to fuel political polarization and manipulate public perception, but its limited reach and rapid debunking minimized broader societal instability.