The recent case of a driver fined for allegedly using a mobile device on Gungahlin Drive in Canberra highlights a critical issue with AI traffic cameras and their potential to unfairly penalize innocent drivers. This incident, where a former public servant with AI expertise identified an optical illusion, underscores the limitations of the technology and the human review process. The driver's open hand, moving from the gear shift to the steering wheel, was misinterpreted as holding a phone, leading to a $548 fine and three demerit points. This case is not an isolated incident, as other drivers have been fined for holding a sushi roll or a wallet, or for wearing a dark jumper that obscured the seatbelt. The problem lies in the execution of the technology, where the AI system can misinterpret benign actions, and the human review process fails to correct these initial mistakes. The burden of proof is then unfairly placed on the driver, who must invest time, stress, and resources to prove their innocence. This situation erodes public trust in road safety laws, as drivers encounter inflexible and unresponsive bureaucracies. The authorities must take responsibility for these systemic failures and audit the algorithms powering the traffic cameras. Technology providers should refine the AI to distinguish between a mobile phone and everyday items or actions. Additionally, infringement review teams need better training and the authority to dismiss false positives without prolonged appeals. It is unacceptable for a just legal system to force innocent drivers to spend months proving a computer wrong. This issue raises a deeper question about the balance between technological advancement and human oversight in law enforcement, and the need for a more transparent and accountable process to ensure public trust and compliance.
