SearchGPT Reportedly Misleads Users with Incorrect Festival Dates in Demo

OpenAI’s prototype AI tool, SearchGPT, provided incorrect dates for An Appalachian Summer Festival in Boone, North Carolina during a demonstration video. The AI listed dates that were incorrect, potentially misleading users planning to attend the event, but also harming the reputation of OpenAI as the incident occurred during a high-profile event.

Risk classification

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

The SearchGPT prototype generated discrete false facts, specifically incorrect dates for a music festival and an incorrect speaker list, which represents the inadvertent generation of misleading information.

Additional risk subdomains

• 7.3 Lack of capability or robustness: The hallucinations stem from the fundamental technical limitations of prediction-based large language models failing to perform reliably.

Causal factors

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

The hallucination occurred in a publicly released promotional demonstration video and was experienced by early testers who had been granted access to the deployed prototype.

EU AI Act risk tier

• Risk tier: 3 Limited Risk

Limited Risk: SearchGPT is a conversational search chatbot, which falls under Risk Level 3 (Limited Risk) as a chatbot requiring transparency about its AI nature.

AI system and alleged parties

• AI system: SearchGPT prototype (OpenAI)
• AI purpose: Content Search; Question Answering
• Behaviour type: Assistant
• Alleged developer: OpenAI
• Alleged deployer: OpenAI
• Alleged harmed parties: OpenAI, An Appalachian Summer Festival attendees

Harm severity

Severity ratings are based on this scale.

Highest direct severity in any category: Negligible. 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 Negligible, indirect Negligible
• 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 Negligible
• Epistemic: direct Negligible, indirect Negligible
• Child sexual exploitation and abuse: direct Negligible, indirect Negligible

People affected

• Occurrences reported: 1
• People reportedly exposed: 10000

Potential causes

Management

• Rushed Market Entry: Rushing to compete with Google and Perplexity despite technical limitations.
• Lack of Clear Business Model: Rushing to search because AI companies lack proven business models.

Technology

• Next-Word Prediction Flaws: Models predict words instead of understanding facts, causing hallucinations.
• Inherent Hallucination Tendency: Generative AI models are fundamentally prone to making up false information.

Data Inputs

• Lack of Specialized Data Access: Lacks Google's expensive, specialized data like flight or local business info.
• Poor Web Scraping and Synthesis: Tool struggles to accurately copy-paste or synthesize data from web sources.

Human Factors

• Unwarranted User Trust: Users expect breakthrough results but face confusing or unhelpful interfaces.

Process and Methods

• Hasty Prototype Release: Releasing a highly anticipated prototype with known factual inaccuracies.
• Iterative Error-Prone Cycle: Releasing flawed products to get feedback instead of robust internal testing.

Information quality

• Classification confidence: High
• Reason for confidence: The reports provide clear, consistent, and detailed accounts of the specific errors made by SearchGPT, including the exact dates hallucinated and the actual dates verified with the festival box office. The technical nature of the system as a prototype search chatbot is explicitly stated.
• Ambiguities identified: None of significance; the reports are highly consistent regarding the specific errors.
• Alternative interpretations: The incident could be viewed purely as a minor product development bug rather than a safety risk, though the systemic nature of LLM hallucinations makes it relevant to epistemic safety.

OpenAI's SearchGPT prototype hallucinated incorrect dates for a local music festival and speaker details during a promotional demonstration. This incident represents a minor, unintentional technical error in a pre-release consumer product with negligible national security implications.

• Overall national security impact: Negligible
• Scope: Single nation
• Primary target: No clear primary
• Alleged perpetrator: Unknown

Threat characteristics

• Imminence: Long-term. The incident highlights a persistent technical limitation of LLMs (hallucinations) rather than an active national security crisis.
• Autonomy: Human-supervised. The AI system autonomously retrieves and synthesizes information, but operates under user prompting and oversight.
• Novelty: Established threat. AI hallucinations and factual errors in search/LLM outputs are well-documented, established technical limitations.

Impact by dimension

• Physical security: Negligible. The incident involves a search engine prototype hallucinating music festival dates, posing no threat to physical systems, critical infrastructure, or human safety.
• Information security: Negligible. The hallucination was an unintentional technical error in a prototype demo, not a coordinated information warfare campaign or compromise of intelligence capabilities.
• Sovereignty: Negligible. The error concerns local community event details and has no impact on state authority, electoral systems, or government decision-making processes.
• Economic security: Negligible. While demonstrating limitations in AI search technology, the incident did not result in theft of strategic technology, financial system attacks, or critical supply chain disruptions.
• Societal stability: Negligible. Incorrect festival dates and speaker listings do not threaten social cohesion, civil liberties, or population safety, nor do they implement mass surveillance.