![\"ai](\"https:\/\/www.newsbeep.com\/au\/wp-content\/uploads\/2026\/02\/ai-traffic-camera-errors-531164175.jpg\")
The Finer Details<\/p>\nA criminal lawyer has warned that AI traffic cameras used nationwide are issuing incorrect infringement noticesThe Key Details<\/p>\n
AI cameras now active in all Australian states<\/p>\n
Lawyer reports repeated misidentification of objects as mobile phones<\/p>\n
Authorities claim human review occurs before an infringement notice is issued<\/p>\n
Court challenges to AI-based fines are increasing<\/p>\n
Drivers can request an internal review via state-based roads authority before heading to court<\/p>\n
The Finer Details<\/p>\n
Astor Legal\u2019s criminal lawyer Avinash Singh<\/a> claims the new technology, which was switched on across the country mid-last year, is issuing false detections, particularly for alleged mobile phone offences<\/a>.<\/p>\n That\u2019s a crime that could land you a fine of more than $1000 depending on the state<\/a>.<\/p>\n Singh said the AI tech is failing to reliably distinguish between phones and common objects.<\/p>\n \u201cAI cameras are notoriously unreliable at detecting mobile phone use in particular. This is because the technology often mistakes other objects for mobile phone cameras,\u201d Singh said.<\/p>\n \u201cSome examples from recent cases we have acted in include wallets, glasses cases and battery packs.\u201d<\/p>\n State transport departments<\/a> state that when AI flags a potential offence, the image is then subject to human review and \u201cchecked and verified by qualified independent officers\u201d.<\/p>\n However, Singh believes there is \u201cno way of confirming this,\u201d adding that it \u201cmakes the issue all the more troubling given the vast number of false detections that are issued.\u201d<\/p>\n According to Singh, AI-generated infringement notices are being challenged more frequently but the outcomes often depend on whether the prosecution can prove the allegation beyond reasonable doubt using the AI-captured image.<\/p>\n \u201cTo challenge an AI-based detection, the fine first needs to be elected to court without being paid,\u201d Singh said.<\/p>\n \u201cThe case will then be set for a hearing, where the driver should have a criminal defence lawyer appear for them to present evidence and argue why the prosecution cannot prove beyond a reasonable doubt that the AI-based image shows a mobile phone.\u201d<\/p>\n However, drivers can request a review from their state-based roads authority, such as Vicroads or Transport for NSW.<\/p>\n According to the NSW road authority<\/a>, you can request a review of a fine (fine notice, fine reminder notice or overdue fine) if:<\/p>\n you believe a mistake has been made, or<\/p>\n there were special circumstances that led to the offence<\/p>\n The website states you could be let off with a caution, the fine may be cancelled, or it might stick after the review.<\/p>\n \u201cRevenue NSW will consider your personal situation and any special circumstances that resulted in you receiving the fine,\u201d reads the website.<\/p>\n In Victoria, it\u2019s a similar process \u2013 see the Fines Victoria website<\/a> for all the details on requesting a infringement review.<\/p>\n We asked around the office and got a telling example: one colleague’s brother was fined for “holding his phone” on the Hume Highway in NSW. However, he was holding his sunglasses case, not a phone.<\/p>\n The AI camera flagged it as an offence, but when he reviewed the images on Revenue NSW’s website, he clearly identified that it was not his phone. He successfully contested the circa-$400 fine by requesting a review and submitting photos of the sunglasses case along with his car’s smartphone holder mounted to the air vent \u2013 which shows where the phone is located.<\/p>\n The fine was dropped.<\/p>\n According to Singh, not everyone will achieve the same result.<\/p>\n “A review can be requested but this is usually only successful if the driver has a clean driving history for the past 10 years,” he said.<\/p>\n With courts seeing more challenges to AI-based detections, state authorities may face pressure to provide clearer information about review processes and error rates.<\/p>\n Increased scrutiny of the technology is expected as more drivers dispute the accuracy of their fines.<\/p>\n AI \u2018slop\u2019 more broadly has become a problem in today\u2019s digital society and has created significant public distrust around AI. Artificial intelligence\u2019s ability to identify objects and make decisions<\/a> is clearly far from perfect.<\/p>\n![\"ai](\"https:\/\/www.newsbeep.com\/au\/wp-content\/uploads\/2026\/02\/ai-traffic-camera-errors-476185162.jpg\")
<\/p>\n![\"\"\/](\"https:\/\/www.newsbeep.com\/au\/wp-content\/uploads\/2026\/02\/avinash-singh-headshot-1.jpg\")
Avinash Singh<\/p>\n![\"ai](\"https:\/\/www.newsbeep.com\/au\/wp-content\/uploads\/2026\/02\/ai-traffic-camera-errors-1146114591.jpg\")
The Road Ahead<\/p>\n