Qld AI traffic cameras flagged over missing ethical checks

By
Follow google news

Part of mobile phone and seatbelt technology program.

An image-recognition AI system used to assess millions of photos captured of drivers in Queensland has come under scrutiny for not adequately managing potential ethical risks.

Qld AI traffic cameras flagged over missing ethical checks

The system, part of the state's mobile phone and seatbelt technology (MPST) program, uses AI-supported 'heads-up' cameras developed by Acusensus to detect mobile phone use and seatbelt non-compliance.

It was first deployed by the Department of Transport and Main Roads (TMR) in July 2021.

While operationally effective, reducing the number of cases requiring human review by 98 percent, the system was flagged in a new audit for potentially falling short of ethical expectations, particularly in light of the Queensland Government’s AI Ethics Framework released in September 2024.

According to the audit [pdf], 114,000 fines were issued in 2024 following human reviews of AI-flagged cases by both the vendor and the Queensland Revenue Office, which administers infringements.

“The MPST program uses image-recognition AI to detect driving offences, which introduces a range of ethical risks,” the auditor’s report said.

“TMR has not yet undertaken a full ethical risk assessment as required by the Queensland government’s AI governance policy.

"This means it does not know whether all ethical risks for the MPST program are identified and managed.”

The report, however, noted that the MPST program was introduced before Queensland issued its AI governance policy.

It also acknowledged that TMR “has implemented controls to support the system reliability and accuracy, protect privacy, enable fair outcomes in the fine adjudication process, and manage its contractual arrangements”.

TMR responded to the audit saying it is currently building a framework to provide “centralised visibility of its AI systems” over the next 12 months and also intends to use the state's FAlRA -foundational artificial intelligence risk assessment - framework by the end of 2025.

TMR is currently in the middle of a program to expand the number of MPST units it operates, having signed a $27.4 million contract extension with the ASX-listed Acusensus in December last year

On the whole, the audit raises concerns that TMR “lacks full visibility” over its AI systems in use,

This extends to its adoption of the state government’s own AI assistant, QChat.

According to the audit, TMR has not yet configured any entity-specific prompts for QChat.

“Using this safeguard could improve the accuracy of responses from QChat and reduce the risk of users receiving misleading guidance,” the audit stated.

The department responded by saying it planned to introduce monitoring procedures and implement them by December 2025, as well as improve internal AI literacy.

Got a news tip for our journalists? Share it with us anonymously here.
Copyright © iTnews.com.au . All rights reserved.
Tags:

Most Read Articles

CBA builds Lumos, an AI-driven accelerator to prepare applications for cloud

CBA builds Lumos, an AI-driven accelerator to prepare applications for cloud

NSW Education AI tool set to launch for students from October

NSW Education AI tool set to launch for students from October

Kmart Australia's use of facial recognition breached privacy

Kmart Australia's use of facial recognition breached privacy

Woodside Energy sets up a unified enterprise data platform

Woodside Energy sets up a unified enterprise data platform

Log In

  |  Forgot your password?