Purportedly AI-Generated Travel Information Reportedly Misled Tourists into Seeking Nonexistent 'Sacred Canyon of Humantay' in Peru

Two tourists in Peru reportedly relied on AI-generated travel information describing a nonexistent destination called the "Sacred Canyon of Humantay." One tourist had paid nearly $160 to reach a rural road near Mollepata without a guide or actual destination. Trek operator Miguel Ángel Góngora Meza intervened after seeing the purportedly AI-generated itinerary and warned that such misinformation could expose travelers to serious risks in high-altitude Andean terrain.

Risk classification

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

The AI systems generated false and misleading travel information, leading users to form inaccurate beliefs and make decisions that resulted in financial loss and physical stranding.

Additional risk subdomains

• 7.3 Lack of capability or robustness: The core technical failure is the AI models' lack of capability to prevent hallucinations and verify geographical facts.
• 5.1 Overreliance and unsafe use: Users relied on general-purpose chatbots for safety-critical planning, such as mountain hiking schedules and high-altitude routes.

Causal factors

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

The incidents were caused by deployed AI systems generating unintended hallucinations and false information during user interactions.

EU AI Act risk tier

• Risk tier: 3 Limited Risk

Limited Risk: The systems described are chatbots (ChatGPT, Copilot, Gemini, Layla) used for generating travel itineraries, which fall under the category of chatbots and AI-generated content requiring transparency.

AI system and alleged parties

• AI system: ChatGPT, Google Gemini, Layla, Microsoft Copilot, unspecified (Google, Microsoft, OpenAI)
• AI purpose: Chatbot; Navigation Assistant
• Behaviour type: Assistant
• Alleged developer: OpenAI, Large language model developers, Chatbot developers
• Alleged deployer: Travelers, Tourists seeking the supposed Sacred Canyon of Humantay, tourists, Chatbot users
• Alleged harmed parties: Travelers, tourists, OpenAI users, Epistemic integrity, Consumers, Chatbot users

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

Epistemic

Reported: The report explicitly describes epistemic harm through AI-driven misinformation and fabrication of geographic facts.

Directly caused: AI systems generated fabricated destinations like the 'Sacred Canyon of Humantay', non-existent cable cars in Malaysia, and incorrect ropeway operating hours in Japan, misleading multiple users.

Inferred additional harm: Widespread erosion of trust in online travel information and maps due to the proliferation of AI-generated travel content.

People affected

• Occurrences reported: 4
• People reportedly harmed: 7
• People reportedly exposed: 7

Potential causes

Management

• Prioritizing Engagement Over Safety: Platforms deploy generative tools without ensuring factual accuracy for safety.

Technology

• Inherent LLM Hallucinations: LLMs generate words statistically, leading to invented facts or destinations.
• Lack of Physical World Understanding: AI cannot distinguish physical realities like mountain climbs from city walks.
• Outdated Information Processing: AI provided outdated ropeway closing times, leaving tourists stranded.

Data Inputs

• Statistical Word Association: AI combined regional terms to invent the non-existent Humantay Canyon.
• AI-Generated Media on Social Media: TikTok videos generated by AI created realistic but fake tourist attractions.

Human Factors

• Over-reliance on AI Planning: Travelers trusted AI itineraries without verifying safety-critical details.
• Lack of Verification by Users: Users failed to double-check AI advice before hiking in remote areas.

Process and Methods

• Lack of Fact-Checking Mechanisms: AI platforms lack real-time verification to filter out false travel details.
• No Distinction in Output Presentation: AI presents hallucinations and facts with the same confident tone.

Regulatory Environment

• Absence of AI Watermarking Rules: Lack of standard labels makes it hard to detect AI-generated travel media.
• Lack of Prevention Regulations: Current regulatory focus is on mitigation rather than preventing AI errors.

Information quality

• Classification confidence: High
• Reason for confidence: The reports provide clear, consistent, and detailed accounts of multiple specific incidents across different countries, supported by quotes from affected travelers, local experts, and academic sources.
• Ambiguities identified: The exact AI models used in the Malaysia TikTok incident are not fully specified, and the exact financial losses for all occurrences are not fully quantified.
• Alternative interpretations: The incidents could be interpreted purely as user error due to overreliance, but the primary failure is the AI's generation of false information.

Commercial AI chatbots generated fabricated travel itineraries and geographic details, leading to minor financial losses and localized stranding of tourists in countries like Japan and Peru. While highlighting safety concerns regarding user overreliance on LLMs for navigation, the incident has negligible national security implications and presents no threat to critical infrastructure, sovereignty, or state stability.

• Overall national security impact: Negligible
• Scope: Multiple nations
• Primary target: No clear primary
• Other affected: Japan, Peru, Malaysia, Italy
• Alleged perpetrator: Unknown

Threat characteristics

• Imminence: Long-term. Represents an ongoing strategic concern regarding the reliability of LLMs rather than an active national security crisis.
• Autonomy: Human-controlled. The AI systems act as assistants generating information, but humans retain full control over whether to execute the suggested plans.
• Novelty: Established threat. LLM hallucinations and generation of plausible-sounding but false information are well-documented and established technical limitations.

Impact by dimension

• Physical security: Negligible. While tourists faced minor physical risks such as being stranded on a mountain or left in high-altitude areas, there was no threat to critical infrastructure, military systems, or national physical security.
• Information security: Negligible. The incident involves unintentional commercial LLM hallucinations rather than coordinated state-sponsored information warfare, intelligence compromise, or systematic disinformation.
• Sovereignty: Negligible. No government functions, electoral systems, border controls, or sovereign decision-making processes were compromised or targeted in this incident.
• Economic security: Negligible. Financial impacts were limited to minor, individual-level travel expenses with no systemic threat to financial institutions, strategic industries, or technological competitiveness.
• Societal stability: Negligible. The incident caused minor personal frustration and localized travel disruptions, but posed no threat to societal cohesion, civil liberties, or population-scale stability.