Evelyn Trueman, a 79-year-old pensioner from Brisbane, is set to represent herself against the Queensland government in a legal battle over a $1,078 fine generated by the state's Camera Detected Offence Program (CDOP). Trueman, who has never been to court before, alleges that the AI system erroneously identified her vehicle as violating seatbelt regulations.
Trueman's case revolves around a series of images provided by the government, where she contends that the car depicted is not hers. Despite driving along the Pacific Motorway with her daughter on the same day, she insists that the passenger seat occupant in the images is not her. Notably, she points to a reflection of her upside-down number plate in the car's windscreen as evidence.
While experts, such as Griffith University quantum physicist A/Prof Erik Streed, find it unlikely for a number plate to be reflected upside down in such circumstances, Trueman argues that the burden of proof should lie with the government. The Queensland Revenue Office (QRO) states that images issued on infringement notices are unaltered, maintaining the accuracy of the AI system.
The CDOP, initiated in 2020 to detect mobile phone infringements and expanded to seatbelt offenses in 2021, uses AI to filter images captured by road safety cameras. Despite authorities emphasizing the involvement of trained officers in reviewing AI-identified images before issuing fines, concerns have arisen regarding the system's accuracy.
The widespread use of these cameras, coupled with substantial increases in fines, has resulted in a nearly 70% surge in traffic camera revenue in the last financial year, reaching $466 million. Several individuals, including Trueman and Gavin Storkey, a Gold Coast small business owner, are challenging their fines, expressing concerns about the reliability of the system.
Critics argue that the rapid implementation of AI in law enforcement and revenue collection should be approached with caution, emphasizing the potential for errors. While AI offers scalability and predictive capabilities, the Queensland Council for Civil Liberties President, Michael Cope, calls for public review and transparency in the deployment of such technologies, citing the impact on people's lives.
As Queensland pensioners like Trueman navigate the legal system alone, questions surrounding the accountability and oversight of AI-generated fines persist. The broader implications of relying on automated systems for critical decisions in law enforcement underscore the need for careful consideration and public scrutiny.
