Deepfakes of Deceased Indian Politicians for Election Campaigning Are Increasingly Being Deployed

In the lead-up to India's 2024 general elections, AI technology was used to create deepfake videos of deceased politicians, such as M. Karunanidhi and J. Jayalalithaa, aiming to influence voter behavior and campaign strategies. These AI-generated appearances are contributing to the erosion of trust in democratic processes and media discourse.

Risk classification

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

Political campaigns are intentionally using AI-generated deepfakes of deceased and living politicians to conduct coordinated, large-scale influence campaigns aimed at manipulating public opinion and political processes.

Additional risk subdomains

• 3.1 False or misleading information: The AI systems generate synthetic speeches that actively ascribe opinions and statements to deceased politicians, creating misleading representations of their views.

Causal factors

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

The incident is caused by the intentional decisions of human political campaigns and consultants to deploy AI deepfakes to influence voters post-deployment.

EU AI Act risk tier

• Risk tier: 3 Limited Risk

Limited Risk: The reports describe the creation and dissemination of AI-generated content such as deepfakes, which are subject to transparency and labeling obligations under the EU AI Act.

AI system and alleged parties

• AI system: ElevenLabs, Muonium AI, The Indian Deepfaker (ElevenLabs, Muonium)
• AI purpose: Deepfake Video Generation; Voice Generation
• Behaviour type: Tool
• Alleged developer: The Indian Deepfaker, Muonium
• Alleged deployer: Various Indian political parties, Dravida Munnetra Kazhagam, DMK
• Alleged harmed parties: Media discourse, M. Karunanidhi, J. Jayalalithaa, Indian voters, Indian electorate, Electoral integrity, 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 Substantial
• Child sexual exploitation and abuse: direct Negligible, indirect Negligible

Malicious content

Reported: The report explicitly describes the creation and spread of malicious synthetic content targeting and mocking political opponents.

Directly caused: An AI-generated video of arrested politician Arvind Kejriwal singing behind bars and a digitally altered video of Muslim politician Asaduddin Owaisi singing Hindu devotional songs were created and shared on social media platforms.

Inferred additional harm: It is highly likely that further malicious political deepfakes are being generated and shared across encrypted messaging networks like WhatsApp to target specific voter groups.

Democracy

Reported: The report explicitly describes the erosion of democratic norms caused directly and indirectly by the incident.

Directly caused: The communications minister stated that deepfakes are 'a serious threat to democracy and social institutions', and experts warn they threaten 'what we can and cannot believe'.

Indirectly caused: The use of resurrected dead leaders and synthetic videos to influence voters undermines the integrity of the 2024 general elections.

Inferred additional harm: Widespread deployment of unverified synthetic media could permanently erode public trust in democratic processes and election outcomes.

Epistemic

Reported: The report explicitly describes epistemic harm caused directly and indirectly by the incident through the creation of synthetic audio and video that blurs reality.

Directly caused: AI systems cloned the voices of deceased leaders like Karunanidhi and Jayalalithaa to deliver political messages they never spoke, making voters struggle to decipher reality from fiction.

Indirectly caused: The proliferation of highly realistic synthetic media pollutes the information ecosystem, making it difficult for citizens to trust genuine political communications.

Inferred additional harm: The widespread availability of cheap voice cloning and deepfakes will likely lead to a broader 'liar's dividend' where real political events are dismissed as synthetic.

People affected

• Occurrences reported: 5
• People reportedly harmed: 4
• People reportedly exposed: 500000

Potential causes

Management

• Prioritizing Low Cost over Ethics: Parties choose cheap digital resurrection over expensive physical rallies.
• Commercialization of AI Tools: Vendors selling cheap voice-cloning services without strict ethical oversight.

Technology

• Imperfect Lip Sync and Visuals: Unavailability of high-quality video datasets led to imperfect visual outputs.
• Rapidly Advancing Voice Cloning: Advanced AI tools allow highly realistic replication of deceased leaders' voices.

Data Inputs

• Sourcing Public Internet Data: Low-quality public video data was used to train the likeness of the leader.
• Cadre-Supplied Scripting: Scripts for resurrected leaders were supplied and vetted by party personnel.

Human Factors

• Low Digital Literacy: A large part of the public struggles to distinguish AI media from reality.
• Voter Cognitive Bias: Voters only tend to fact-check videos that do not align with their beliefs.
• Sentiment Exploitation: Parties leverage emotional attachment to dead leaders to sway voter opinions.

Process and Methods

• Inadequate Ethical Vetting: No standardized ethical vetting process exists for political consultants.
• Mass Dissemination via WhatsApp: Unchecked distribution channels allow rapid spread of synthetic media.

Regulatory Environment

• No Rights for Deceased People: Indian law lacks legal framework protecting deceased individuals' voice rights.
• Absence of AI Campaign Regulations: Lack of clear regulatory guidelines on the use of AI in election campaigning.

Information quality

• Classification confidence: High
• Reason for confidence: The reports provide clear, detailed, and consistent accounts of how political parties in India are using AI-generated deepfakes of living and deceased politicians. Multiple sources, including AI creators, political consultants, and tech policy experts, are quoted, providing high credibility.
• Ambiguities identified: None of significance; the technical methods (voice cloning, generative video) and deployment channels (WhatsApp, public screens) are well-described.
• Alternative interpretations: Some might view this as standard political campaigning/satire rather than a malicious disinformation threat, though the consensus in the text leans towards democratic risk.

Political campaigns in India are deploying advanced AI-generated deepfakes and voice clones of living and deceased politicians to influence voters in the 2024 elections. This widespread use of synthetic media poses a substantial threat to information security, democratic processes, and societal trust, representing an evolved capability in domestic political influence operations.

• Overall national security impact: Substantial
• Scope: Single nation
• Primary target: India
• Alleged perpetrator: Indian political campaigns

Threat characteristics

• Imminence: Near-term. The threat is active and developing during the 2024 general election cycle, requiring ongoing monitoring and response.
• Autonomy: Human-controlled. AI systems are utilized as tools by human operators who dictate the scripts, targets, and distribution of the generated media.
• Novelty: Evolved capability. While political deepfakes are established, the systematic resurrection of deceased leaders at scale represents a significant evolution in campaign tactics.

Impact by dimension

• Physical security: Negligible. No physical threats, kinetic attacks, or critical infrastructure manipulation were reported in connection with these deepfake campaigns.
• Information security: Substantial. Systematic deployment of realistic voice clones and deepfakes across social media platforms represents a substantial information operations threat to the election information ecosystem.
• Sovereignty: Substantial. The integrity of democratic processes and national elections is challenged by the widespread dissemination of synthetic media designed to influence voter behavior.
• Economic security: Negligible. No significant threats to financial systems, critical supply chains, or strategic technological advantages were identified.
• Societal stability: Substantial. Large-scale synthetic manipulation risks eroding public trust in democratic institutions, social cohesion, and the shared perception of reality.