‘Echo chamber’: Dangerous AI trend on the rise among AussiesWhile AI bots like ChatGPT offer free advice at the touch of a button, psychologists have warned the increasing trend could spell disaster for vulnerable Australians.Alexandra Feiam5 min readOctober 5, 2025 - 1:48PMNewsWireIt’s been three years since ChatGPT hit the general public, and a staggering number of Australians are turning to it for their mental health check-ups.AUniversity of Sydneystudy found that about one-in-10 Australians use the AI platform for health-related questions.Be it a sore stomach, unusual-looking mole or a chronic headache, it’s seen as an easy way to determine whether an ailment is worth visiting a doctor.What’s more, many are turning to the chatbot as a therapist, sharing their deepest secrets in place of a registered psychologist.It’s a trend that’s on the rise in Australia – and it has psychologists worried.A worrying increase of Australians are turning to AI chatbots for mental health support Picture: NewsWire / Christian GillesConvenient, fast and often free, it offers people a chance to ask questions about their mental health without having to wait – or pay – to see a psychologist.About20 per centof TikTok users have admitted to using AI for therapy, too.“We know many people are using them, usually due to barriers such as cost, stigma or access,” clinical psychologist Katie Kjelsaas told NewsWire.“AI platforms are accessible, immediate and anonymous – the ease and privacy of access may result in many people giving them a try as an alternative to therapy without fully considering the implications of this.”Dr Kjelsaas said AI platforms such as ChatGPT posed a serious risk to users.“I can see AI chatbots could serve a purpose here in providing limited interim relief to people with low-level distress, but the issue is with the total inability of AI to perceive/discern the distress level or need of a user,” she said.“This makes seeking help from AI – as an alternative to seeking help from a skilled professional – downright dangerous, especially given many people experiencing acute mental distress lack the capacity to accurately assess their own symptoms or distress level, let alone report on this in a way that allows a generative AI platform to reach accurate conclusions.”AI bots like ChatGPT are not registered or regulated psychologists. Picture: NewsWire / Christian GillesIs ChatGPT a good alternative for therapy?While it may seem like a cheap and quick fix to seeing a registered therapist, AI platforms like ChatGPT have a long list of downsides.“AI platforms may be able to offer general advice in some cases (although we know they also frequently make significant errors when doing this), but they cannot diagnose, formulate or personalise treatment the way a psychologist can,” Dr Kjelsaas said.“You do not see a doctor to have them regurgitate information about your symptoms. You see a doctor to have them sort, synthesise and personalise the relevant information about your symptoms to you so they can apply this in a way that supports your healing.”Australian Association of Psychology president Sahra O’Doherty said the bots were “designed to summarise” or “mirror” what a user fed it rather than basing its answers on “correct or evidence-based information”.“It can learn about you and your style of writing, but it will not offer new, alternative or challenging perspectives unless the user specifically asks it to, which many people do not know how to do,” she told NewsWire.Even OpenAI – which owns ChatGPT – agrees.“ChatGPT is not a substitute for professional care and does not provide medical, psychiatric, or crisis intervention advice,” the bot told NewsWire.“Users are encouraged to use it as a tool for general information or supportive conversation but not for diagnosis or treatment.”Psychologists warned that chatbots could regurgitate misinformation to vulnerable users. Picture: NewsWire / Christian GillesAre AI platforms regulated?First and foremost, AI platforms like ChatGPT are not registered or trained psychologists. With no regulation or accountability, an AI platform can easily spread misinformation or misinterpret what it’s being fed.“Psychologists are accountable to AHPRA under Australian law,” Dr Kjelsaas said.When asked about its regulation, ChatGPT told NewsWire that it was “designed with clear guardrails against providing specific medical or therapeutic advice” and only offers “general information” to its user.“If a user asks for medical or mental health guidance, the model typically responds with general information and reminds users to consult qualified professionals for actual diagnosis or support,” the AI chatbot said.“Where possible, links to mental health resources (such as crisis helplines) are provided, especially if the conversation implies someone may be at risk of harm.“For teens, parental controls can also trigger safety notifications to a guardian if self-harm is detected.”AI can often get things wrong, even without the user knowing. Picture: NewsWire / Christian GillesCan AI bots get it wrong?Putting it simply – yes.Ms O’Doherty said AI bots like ChatGPT had frequently reported “inaccurate data and source information from all over the internet”, often “without discerning whether the source is accurate or reliable”.“There have also been numerous cases of AI generating false information (making things up) instead of providing factual information,” she said.If mental health concerns are getting worse, she said there was a “real risk” of delayed access to appropriate care, which could make problems and symptoms even worse.The bot told NewsWire that its responses were generated “based on patterns in data and may not always be accurate or complete”.“Users are encouraged to verify any important information, especially regarding health. If ChatGPT produces potentially harmful or incorrect content, users can report the chat for review,” it said.Although it is not regulated like a trained psychologist, the bot said it had reporting tools and “guardrails” built into its app.“OpenAI uses automated systems and a dedicated human team to investigate and take action, including removing or blocking problematic responses, restricting accounts in serious cases, and updating guardrails as needed,” it said.AI bots also store the user’s information. Picture: NewsWire / Christian GillesCan AI bots be used for good in the mental health space?Despite the laundry list of downsides to relying on AI chatbots for mental health checks, Dr Kjelsaas said there was some benefit in using them.“AI platforms are accessible, immediate and anonymous – the ease and privacy of access may result in many people giving them a try as an alternative to therapy without fully considering the implications of this,” she said.Ms O’Doherty said using AI chatbots could be “incredibly useful” when used as a “supplement to therapy”.“It was designed as a journalling and summarising tool, which it can do very well,” she said.“It can pick up on patterns in the information provided to it and reflect back what it has learned about the user. This can be great when we are going through a period of self-reflection, but it does not provide a way out of the echo chamber unless specifically directed to.”Experts have urged the government to increase funding for mental health support. Picture: NewsWire / Christian GillesWhat can people do instead?With more people turning to AI platforms for a quick and cheap way to access mental health, Ms O’Doherty has urged the government to make mental health services more accessible to the general public.“The main factor seeming to drive up the popularity of ChatGPT and AI as alternatives to psychology is the cost of mental health care,” she said.“The government needs to see this as a risk issue and make psychology services more accessible and affordable for patients.”Dr Kjelsaas said people should avoid using AI platforms for “crisis or emergency support, asking for diagnosis, asking for personal advice or providing personal information they would not want stored” and instead seek professional help.“I hope AI will become safer as the technology is refined, but for now I believe it poses a significant risk of harm to those who use it as an alternative to therapy in the absence of support and guidance from a trained mental health professional,” she said.More related storiesAt WorkPsychs win big pay bump after wild battleDoctors in one state have won a 20 per cent pay bump following a wild battle that involved mass resignations and fears of a breakdown in services.Read moreQLDFallen Block star’s candid health updateFormer reality TV star Suzi Taylor has shared a major personal update celebrating a health milestone, after years of being in and out of the courts on various petty offences.Read moreNationalDaylight saving: Everything you need to knowDaylight saving has begun, giving about 18 million Australians an extra hour of afternoon daylight.Read more