‘Doesn’t actually know’: The controversial tool Aussies are using to fix their cars
Australian drivers are turning to AI to diagnose and fix their cars. But are the pocket savings worth the risk? According to experts, it's not worth the headache.
Turn on any smart device, and virtually every feature and social media app nowadays has some form of Artificial Intelligence. As AI becomes more ubiquitous across the world, the high-tech tool has drawn controversy across different industries, and automotive is no exception.
A new survey conducted by the Royal Automotive Club of Queensland (RACQ) – the state’s peak motoring body – has revealed that some Australians are turning to AI to diagnose car problems and DIY repair their cars.
The motoring body polled 600 local drivers, with 14 per cent of survey respondents confessing to using AI as a tool for DIY fixes. A further 23 per cent said they seek advice from friends and family instead of seeing a mechanic.
According to survey respondents, economic pressures are a key factor behind relying on AI for repairs, with a third of respondents stating they've delayed or skipped a car service altogether despite knowing the risks.
Approximately 40 per cent admitted they delayed a service by three to six months, while 21 per cent said they’ve postponed it by more than six months.
Though convenient, experts warn that AI isn’t a perfect tool, and when it comes to complex tasks such as car servicing and repairs, it could be a recipe for disaster.
According to RACQ’s principal technical researcher Andrew Kirk, AI tools cannot provide accurate and in-depth solutions.
“While AI tools, internet searches and advice from well-meaning friends can provide general information, they cannot physically inspect a vehicle, accurately diagnose every fault, or verify that a repair has been performed correctly,” he said in a media statement.
Meanwhile, the Victorian Automotive Chamber of Commerce (VACC) chief executive officer Peter Jones said AI engines gather information from numerous sources, which calls their accuracy into question.
“The core problem is that the driver has no way of knowing where the answer came from. Is the model drawing on the OEM repair procedures and specifications for that specific vehicle, or is it making an educated guess – or hallucinating an answer – because it doesn't actually know? A confident answer and a correct answer are not the same thing, and AI gives you no way to tell them apart," Jones told Drive.
The VACC executive said the reliance on AI for car repairs can lead to three practical and costly traps relating to information, equipment and most importantly, interpretation.
“Repairing a vehicle properly requires OEM-based, verified service, repair and diagnostic information specific to that vehicle. There is no guarantee an AI model has it, and no way for the user to verify that it does," he said.
“Even with accurate information, diagnosis requires automotive diagnostic equipment to gather data on the vehicle's condition and narrow down the problem area. Most DIY repairers don't have access to it — so they're guessing, and the AI is guessing with them.
“Modern vehicles are highly interconnected systems. A symptom in one place frequently originates somewhere else entirely. Trained technicians have the knowledge and experience to read that. There's a misconception that vehicle problems always have simple answers. Sometimes they do. Often the cause is complex and isn't straightforward to identify — and may not even be covered in the OEM diagnostic procedure."
The use of AI tools can also jeopardise a driver's warranty on their vehicle.
As Jones explained, “The biggest risk is that AI can give a DIY repairer incorrect information with complete confidence, and that person doesn't have the experience to recognise it's wrong or the equipment to confirm it. A trained technician can sanity-check a bad suggestion. A driver at home generally can't.”
“There's the warranty question, which drivers rarely think about. That [warranty] protection assumes the work is carried out correctly, to manufacturer specifications, using appropriate parts, and is properly documented.
“DIY work performed on the strength of an AI suggestion meets none of those tests. If that work causes damage, a manufacturer is entitled to deny the claim, and the owner wears the repair bill. Following an AI's advice is not a defence."
On a practical level, Drive journalist Zane Dobie asked ChatGPT in March 2025 to troubleshoot a range of common car issues, from flat tyres to faulty ignitions and blown gaskets, with the AI tool unable to properly solve the problems based on basic human feedback.
While complex car repairs are best left to professionals, a spokesperson for the Victorian Automotive Chamber of Commerce (VACC) – the state's peak maintenance body – told Drive in March 2026 that some drivers can perform basic maintenance tasks that can reduce the final servicing bill.
“Common tasks include checking the engine oil level and coolant, checking driving belt condition, checking tyre pressures and/or inspecting tyres for damage, checking or replacing wiper blades, refilling washer bottles, replacing [light] globes and replacing blown fuses,” a VACC spokesperson said.
