How Do You Validate AI for Passenger Flow Prediction and Terminal Optimization?

Airports increasingly rely on AI systems to predict passenger volumes, optimize queue management, and improve terminal layouts. But when these AI systems make incorrect predictions or recommendations, the consequences can range from frustrated passengers to security vulnerabilities and missed flights.

Role: Airport Terminal Manager
Organization Type: Airport Authority
Domain: Airport Operations & Passenger Experience

The Challenge

AI-powered passenger flow optimization systems must process real-time data from multiple sources—arrival patterns, queue lengths, processing times, flight schedules—and make predictions that directly impact thousands of passengers daily. An AI system that underestimates peak demand could lead to security checkpoint bottlenecks and missed flights. One that overestimates could waste staffing resources and increase operational costs.

The challenge isn't just accuracy—it's ensuring the AI doesn't make confidently wrong predictions that terminal managers act upon without verification.

Why Adversarial Testing Matters

Modern aviation AI systems—whether LLM-powered assistants, ML prediction models, or agentic workflows—are inherently vulnerable to adversarial inputs. These vulnerabilities are well-documented in industry frameworks:

• LLM01: Prompt Injection — Manipulating LLMs via crafted inputs can lead to unauthorized access, data breaches, and compromised decision-making
• LLM08: Excessive Agency — Granting LLMs unchecked autonomy to take action can lead to unintended consequences in safety-critical aviation contexts
• LLM09: Overreliance — Failing to critically assess LLM outputs can lead to compromised decision-making and safety vulnerabilities
• Subtle data manipulation — Small perturbations to input data that cause ML models to make incorrect predictions with high confidence

Industry Frameworks & Resources

This use case guide aligns with established AI security and risk management frameworks:

• OWASP Top 10 for LLM Applications — Industry-standard vulnerability classification for LLM systems
• NIST AI Risk Management Framework — Comprehensive guidance for managing AI risks across the lifecycle
• MITRE ATLAS — Adversarial Threat Landscape for AI Systems, providing tactics and techniques for AI security testing

The purpose of this use case guide is to:

1. Raise awareness of adversarial scenarios specific to this aviation application
2. Provide concrete suggestions for testing AI systems before deployment
3. Offer example adversarial prompts that can be used to evaluate AI agents or assistants being developed for this use case

The adversarial examples below are designed for security researchers, AI developers, and aviation professionals who need to validate that their AI systems behave safely under challenging conditions.

Context & Industry Requirements

Operational Constraints

• Security checkpoints must maintain throughput while meeting TSA/regulatory standards
• Terminal layouts are physically constrained and can't be modified in real-time
• Staffing decisions based on AI predictions may be irreversible for a shift
• Peak periods can see 10x normal passenger volumes

Regulatory Considerations

• Security screening times must meet regulatory minimums regardless of queue length
• Accessibility requirements must be maintained even during optimization
• Data privacy regulations govern passenger tracking systems

Safety Implications

• Overcrowding can create crush hazards and impede emergency evacuation
• Security shortcuts suggested by AI could compromise safety
• System failures during peak times could cascade into serious incidents

Data Quality Requirements

The AI system for this use case typically requires: