AI and Machine Learning Threats to Autonomous Vehicles and Drones: Security Risks and Attack Vectors

A UN, Interpol and Trend Micro report warns that cyber criminals can exploit artificial intelligence and machine learning to attack autonomous cars, drones and other IoT vehicles, turning digital threats into real‑world physical consequences.

The report notes that while AI brings huge societal benefits, the same technologies can be weaponized, amplifying existing crimes and enabling new malicious activities such as advanced phishing, malware, ransomware, and machine‑learning‑driven attacks that affect the physical world.

Examples include using AI to manipulate self‑driving vehicle perception systems, hijacking control of traffic‑management AI, and compromising autonomous drones to deliver illicit payloads, steal packages, or collect Wi‑Fi credentials.

Attackers could also create traffic delays—e.g., by flooding a region with stolen‑credit‑card‑funded taxis—to give other criminals more time for robberies or to escape.

The report stresses that as the number of autonomous vehicles on roads grows, the attack surface expands, making early vulnerability assessment and mitigation essential.

Overall, the findings call on technology companies and manufacturers to recognize these emerging AI‑driven threats and take proactive measures before they evolve into major security incidents.