The New South Wales government has raked in more than $400 million in penalties after rolling out AI-assisted road cameras, but the state is now eyeing an automated fines enforcement system in a proposal that was released to industry this week.
Documentation proposed to industry showed that state officials want a system to automate, verify, and adjudicate incidents on New South Wales roads.
The documents also showed the system would improve efficiency by incorporating AI-assisted decision-making, automation, and intelligent alerting.
NSW Secretary for Transport Josh Murray said humans would not be removed from the fines process entirely, claiming the in-market proposal was not about implementing further AI-linked enforcement measures.
“Every seatbelt or mobile phone image adjudicated by AI that is likely to show a potential offence is reviewed and verified by trained and authorised staff before a penalty notice is issued. This process will not change,” Mr Murray said.
“The Request for Proposal (RFP) which Transport currently has in market is not about implementing additional AI-linked enforcement and will not increase offences detected.”
Since 2021, AI-assisted cameras have generated more than $400 million, 7News said.
But the use of AI on Australian roads remains controversial.
About 130 million photos are captured by AI cameras in Australia annually, before AI analyses each image and forwards it to staff.
In 2025, the false positive rate was 3.74 per cent for mobile phone offences alone, with thousands of potential fines overturned by human reviewers.
In the same year, the false positive rate for seatbelt offences was 15.28 per cent.
NRMA spokesperson Peter Khoury said there needs to be a “human element” in the AI-assisted camera process, highlighting how the technology can form misjudgments.
“Given that if you get a fine you need to prove yourself innocent, it is critical that there is a human element, because we know that the technology can make mistakes,” Mr Khoury said.