AI-Generated Images of the Iconic Hollywood Sign Reportedly on Fire Circulating on Social Media

During the January 2025 Southern California wildfires, AI-generated images of the iconic Hollywood Sign being on fire circulated on social media.

Risk classification

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

The incident involved the generation and viral spread of false AI-generated images depicting a landmark on fire, misleading the public during an active emergency.

Causal factors

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

The incident was caused by human users intentionally generating and sharing fake disaster images using a deployed AI system.

EU AI Act risk tier

• Risk tier: 3 Limited Risk

Risk Level 3. Limited Risk: The report describes AI-generated content such as deepfakes or synthetic images of a disaster, which require transparency obligations so users know they are interacting with AI-generated media.

AI system and alleged parties

• AI system: unspecified
• AI purpose: Image Generation; Chatbot
• Behaviour type: Tool
• Alleged developer: Unknown AI image generator technology developers
• Alleged deployer: social media users
• Alleged harmed parties: Southern California residents

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 Negligible, indirect Negligible
• Democracy: direct Negligible, indirect Negligible
• Privacy: direct Negligible, 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 explicitly describes the creation and spread of fake AI-generated photos and videos of the Hollywood Sign on fire.

Directly caused: X's Grok chatbot directly generated realistic images of the landmark burning when prompted by users.

Indirectly caused: These fabricated images were widely shared and went viral across Meta's platforms and X, flooding social media feeds.

Inferred additional harm: It is likely that other unwatermarked variations of the fake images were generated and shared across private messaging networks.

Psychological

Reported: The report explicitly describes emotional distress and vulnerability among residents who saw the fake images during the active wildfire evacuation.

Indirectly caused: Evacuated residents experienced heightened anxiety and confusion, with social media users noting they felt 'raw and vulnerable' and did not need fake images circulating.

Inferred additional harm: It is highly likely that hundreds of local residents facing evacuation experienced temporary panic and acute distress upon seeing the realistic fake images of the landmark burning.

Epistemic

Reported: The report explicitly describes epistemic harm through the viral spread of AI-generated misinformation that confused the public during an active disaster.

Directly caused: Generative AI tools fabricated realistic visual evidence of a non-existent disaster event (the Hollywood sign burning).

Indirectly caused: The viral spread of these images misled social media users and disrupted the shared understanding of the actual wildfire situation.

Inferred additional harm: The incident likely contributes to a broader erosion of public trust in real-time social media updates during emergency situations.

People affected

• Occurrences reported: 1
• People reportedly harmed: 3
• People reportedly exposed: 10000

Potential causes

Management

• Permissive Release Policy: Management released image generation tools without sufficient guardrails.
• Safety De-prioritization: Decisions to pivot away from active fact-checking prioritized other goals.

Technology

• Grok Image Generation: Grok allowed users to generate realistic images of a landmark on fire.
• Removable AI Watermarks: Grok watermarks were easily cropped or missing in viral versions.
• Algorithmic Amplification: Social media algorithms spread the fake burning sign images rapidly.

Data Inputs

• Unrestricted Prompts: Grok accepted prompts to generate disaster imagery during active fires.

Human Factors

• Malicious Prompting: Users intentionally commanded the AI to create deceptive disaster photos.
• Uncritical Sharing: Social media users reposted fake images without verifying authenticity.
• Truth Apathy: Some users reposted fake images while expressing indifference to reality.

Process and Methods

• Delayed Moderation: Platforms added fact-check disclaimers only after images went viral.
• Reduced Fact-Checking: Meta's pivot away from fact-checking slowed down the verification process.

Regulatory Environment

• Lack of AI Standards: No active regulatory standards prevented the generation of disaster fakes.

Information quality

• Classification confidence: High
• Reason for confidence: The reports provide clear, consistent details about the generation of the fake images, the specific AI tool involved (Grok), and the public reaction. The facts are well-verified by fact-checkers and local authorities.

Users leveraged X's Grok AI to generate realistic fake images of the Hollywood Sign on fire during the January 2025 Southern California wildfires. The viral spread of these images caused temporary panic and confusion among evacuating residents and required local officials to issue debunks, illustrating the growing risk of AI-amplified misinformation during public emergencies.

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

Threat characteristics

• Imminence: Long-term. The incident represents an ongoing strategic concern regarding the generation of synthetic media during public emergencies rather than an active crisis.
• Autonomy: Human-controlled. The AI system acted as a tool directly responding to user prompts, with humans initiating the creation and dissemination of the images.
• Novelty: Evolved capability. While crisis misinformation is an established threat, the rapid generation and viral spread of highly realistic AI-generated imagery during an active disaster represents an evolved capability.

Impact by dimension

• Physical security: Negligible. The AI images did not impact physical critical infrastructure or cause kinetic damage; the landmark itself was never in physical danger.
• Information security: Minor. AI-generated images of a burning landmark went viral, causing localized public panic during an active emergency, but lacked indicators of coordinated state-sponsored information warfare.
• Sovereignty: Minor. Local emergency responders and officials had to divert attention to issue public clarifications, slightly disrupting emergency communication functions during an active disaster.
• Economic security: Negligible. No impact on strategic industries, financial systems, or technological competitive advantage was reported.
• Societal stability: Minor. The synthetic images caused temporary psychological distress, anxiety, and confusion among local residents undergoing evacuation, but did not threaten large-scale societal stability.