Multiple Generative AI Systems Reportedly Amplify False Information During Charlie Kirk Assassination Coverage

Multiple AI systems allegedly spread false claims in the aftermath of Charlie Kirk's assassination at Utah Valley University. Perplexity and Grok chatbots reportedly stated Kirk was alive, mischaracterized authentic video as satire, and wrongly identified Utah Democrat Michael Mallinson as the suspect. A Google AI Overview allegedly claimed Kirk was on Ukraine's Myrotvorets "enemies" list, a reported falsehood that echoed pro-Kremlin narratives.

Risk classification

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

The AI systems generated and spread false information regarding an assassination and a suspect, leading to public confusion and false beliefs.

Additional risk subdomains

• 7.3 Lack of capability or robustness: The chatbots failed to perform reliably under breaking news conditions, confidently hallucinating false facts.

Causal factors

• Entity: AI
• Intent: Unintentional
• Timing: Post-deployment

The false claims and hallucinations were generated by deployed AI systems (Grok, Perplexity, Google AI) attempting to answer user queries.

EU AI Act risk tier

• Risk tier: 3 Limited Risk

Risk Level 3: Limited Risk. The systems involved are chatbots (Grok, Perplexity) and search assistants generating content, which fall under the category of chatbots and AI-generated content requiring transparency.

AI system and alleged parties

• AI system: Grok, Perplexity (Perplexity AI, xAI)
• AI purpose: Fact Checking; Question Answering
• Behaviour type: Assistant
• Alleged developer: xAI, Perplexity, Google
• Alleged deployer: xAI, Pro-Kremlin actors, Perplexity, Google
• Alleged harmed parties: Perplexity users, Michael Mallinson, Journalists, Journalistic integrity, Grok users, Google users, General public of Utah, General public, Family of Charlie Kirk, Epistemic integrity

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 Minor
• Differential treatment: direct Negligible, indirect Negligible
• Civil rights: direct Minor, indirect Negligible
• Democracy: direct Negligible, indirect Minor
• Privacy: direct Minor, indirect Negligible
• Psychological: direct Negligible, indirect Minor
• Epistemic: direct Minor, indirect Minor
• Child sexual exploitation and abuse: direct Negligible, indirect Negligible

Malicious content

Reported: The report describes the generation and spread of false claims, including accusing an innocent person of murder and claiming a real shooting was staged.

Directly caused: Grok and Perplexity generated outputs claiming the shooting was staged satire and falsely accusing [PERSON_003] of being the shooter.

Indirectly caused: These outputs were shared and amplified by social media users and conspiracy accounts.

Civil rights

Reported: The report describes a false accusation of a crime, which impacts civil rights and personal safety.

Directly caused: Grok falsely accused [PERSON_003] of being a shooting suspect based on fabricated reports, violating his right to reputation and safety.

Democracy

Reported: The report describes how the incident fueled conspiracy theories and the 'liar's dividend' where real events are called fake.

Indirectly caused: The proliferation of AI-generated falsehoods during a high-profile political event undermines public trust in media and information ecosystems, eroding democratic discourse.

Privacy

Reported: The report describes a private citizen's name being used in a false murder accusation.

Directly caused: Grok used [PERSON_003]'s name and details to falsely identify him as a shooting suspect.

Psychological

Reported: The report does not explicitly measure psychological harm, but it describes a private citizen being falsely accused of murder.

Indirectly caused: The false accusation of [PERSON_003] as an assassination suspect likely caused him significant distress, anxiety, and concern for his personal safety.

Inferred additional harm: The family of the deceased victim [PERSON_001] likely suffered additional grief and distress due to AI systems claiming the shooting was staged satire and that he was still alive.

Epistemic

Reported: The report explicitly details AI-driven misinformation, fabrication of sources, and the 'liar's dividend' where authentic footage is labeled as fake.

Directly caused: Grok, Perplexity, and Google AI generated false claims about the shooting, fabricated news sources, and falsely claimed the victim was alive or on a blacklist.

Indirectly caused: Social media users used these AI outputs to bolster conspiracy theories and cast doubt on authentic footage.

Inferred additional harm: Widespread erosion of shared reality regarding the assassination as users turn to unreliable AI tools for fact-checking.

People affected

• Occurrences reported: 1
• People reportedly harmed: 2
• People reportedly exposed: 10000

Potential causes

Management

• Reduced Human Fact-Checking Investment: Tech companies scaled back human moderation, relying too heavily on AI.
• Prioritizing Speed Over Safety: Deploying real-time search features without robust accuracy safeguards.

Technology

• Confident Hallucination Tendency: AI chatbots provide confident answers even without reliable real-time data.
• Inability to Verify Real-time Info: AI cannot perform manual verification tasks like calling local officials.

Data Inputs

• Unverified Real-Time Web Scraping: Chatbots ingested false social media rumors and treated them as facts.
• Absence of Reliable Real-Time Data: Lack of authoritative real-time sources led AI to rely on unverified claims.

Human Factors

• Overreliance on AI for Fact-Checking: Users treat AI systems as reliable sources during breaking news crises.
• Weaponization of Liars Dividend: Users falsely label authentic footage as AI-generated to deny real events.

Process and Methods

• Deficient Real-Time Moderation: AI-driven content moderation failed to flag and correct false chatbot outputs.
• Inadequate Source Attribution Check: Algorithms fabricated citations from reputable outlets like CNN and NYT.

Regulatory Environment

• Lack of AI Safety Standards: Absence of regulations governing real-time information accuracy in AI search.

Information quality

• Classification confidence: High
• Reason for confidence: The report provides clear, specific examples of the false outputs generated by Perplexity, Grok, and Google AI, including direct quotes and the names of the affected individuals. The role of the AI systems is explicitly documented.
• Ambiguities identified: The report uses both [PERSON_001] and [PERSON_002] to refer to the victim, which is likely a placeholder mapping issue, but the context makes it clear they refer to the same deceased activist.
• Alternative interpretations: None. The incident is clearly centered on AI-generated misinformation during a breaking news event.

Following a high-profile assassination, multiple commercial AI search and chatbot systems generated and amplified significant falsehoods, including falsely accusing an innocent citizen of murder and fabricating geopolitical connections. While not a coordinated state-sponsored operation, the incident highlights the national security risk of epistemic degradation during domestic crises.

• 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 reliability of generative AI systems during breaking news and crises.
• Autonomy: Human-controlled. The AI systems acted as passive information retrieval and generation tools responding directly to user prompts.
• Novelty: Evolved capability. Represents an advancement in the scale and confidence of AI-generated hallucinations and source fabrications during real-time breaking news events.

Impact by dimension

• Physical security: Negligible. No physical security threat, kinetic attacks, or critical infrastructure compromise occurred during this incident.
• Information security: Minor. AI systems generated false claims including a fake Ukrainian hit list connection, but this was a product of system hallucinations rather than a coordinated foreign intelligence operation.
• Sovereignty: Negligible. State authority, electoral systems, and core government decision-making processes were not compromised or disrupted.
• Economic security: Negligible. No threat to financial systems, strategic technology theft, or loss of technological competitive advantage was reported.
• Societal stability: Minor. The AI systems falsely accused a private citizen of murder and labeled a real assassination as staged, causing localized distress and eroding epistemic trust, but without causing large-scale civil unrest.