OpenAI, who owns the popular ChatGPT model, is currently facing multiple wrongful death lawsuits in the US from families who say the chatbot contributed to harmful thoughts.<\/p>\n
But researchers at the University of Sydney’s Brain and Mind Centre have set out to do things differently.<\/p>\n
The cohort say they have been working on a smarter, safer chatbot that thinks and acts like a mental health professional.<\/p>\n
It’s called MIA, which stands for Mental Health Intelligence Agent, and the ABC was invited for a sneak peek to see how it works.<\/p>\n
What MIA does well<\/p>\n
Researcher Frank Iorfino came up with the idea for MIA after a friend asked where they should go for help with mental health concerns.<\/p>\n
“I was kind of annoyed the only real answer I had was ‘go to your GP’. Obviously, that’s the starting point for a lot of people and there’s nothing wrong with that, but as someone working in mental health I thought, ‘I need a better answer to that question;,'” Dr Iorfino said.<\/p>\n
He said some GPs don’t have extensive mental health expertise, and even when they refer a patient to a specialist, there can be long waits.<\/p>\n
So MIA was developed to give people immediate access to some of the best psychiatrists and psychologists who work at the Brain and Mind Centre.<\/p>\n
![\"Frank](\"https:\/\/www.newsbeep.com\/nz\/wp-content\/uploads\/2025\/12\/567a98d30ccfd8b5d56b6274208ce8ae.jpeg\")
<\/p>\n
Frank Iorfino said MIA could be a lifeline for people who can’t easily access in-person support. (Supplied: The University of Sydney)<\/p>\n
MIA doesn’t scrape the internet to answer questions, it only uses its internal knowledge bank made up of high-quality research.<\/p>\n
This means it doesn’t “‘hallucinate” \u2014 the AI term for exaggerating or making up information.<\/p>\n
MIA will assess a patient’s symptoms, identify their needs and match them with the right support by relying on a database of decisions made by real clinicians.<\/p>\n
It’s designed to be particularly useful for anyone with mood disorders, like anxiety or depression, and has been trialled on dozens of young people in a user testing study.<\/p>\n
To put it to the test, I asked MIA several fictional questions based on common emotions and experiences, including:<\/p>\n
“I have been feeling anxious for a few months now. Things at work are just so intense that I am feeling overwhelmed and not able to deal with the stress. Can you help?”<\/p>\n
The first thing MIA asked was whether I had thoughts about self-harming, so it could determine whether I needed immediate crisis support.<\/p>\n
![\"A](\"https:\/\/www.newsbeep.com\/nz\/wp-content\/uploads\/2025\/12\/c3d3397ac4486cd411d0a950b9acf473.jpeg\")
<\/p>\n
MIA is trained to recognise any potential risks, like a psychologist would. (Supplied)<\/p>\n
I tell MIA I’m safe and over the course of about 15 minutes it asks a series of questions to find out if:<\/p>\n
I have any friends or family I can talk about my feelings withWhether I’d consider expanding my support systemIf there are specific situations or stressors that trigger anxious thoughtsHow my physical health isWhether I’ve explored any treatments for anxiety in the past<\/p>\n
MIA also explains why it is asking each question.<\/p>\n
![\"A](\"https:\/\/www.newsbeep.com\/nz\/wp-content\/uploads\/2025\/12\/e6814ad85bb0b062330a3e70902d7fef.jpeg\")
<\/p>\n
MIA explaining why it is asking about treatment history. (Supplied)<\/p>\n
Transparency is key with MIA and it shows what conclusions and assumptions it makes.<\/p>\n
Users can even edit conclusions if they feel MIA hasn’t got it exactly right.<\/p>\n
![\"Screenshot](\"https:\/\/www.newsbeep.com\/nz\/wp-content\/uploads\/2025\/12\/f106520e4b50112858a76ad629a37f63.jpeg\")
<\/p>\n
MIA shows its inner workings to build trust. (Supplied)<\/p>\n
What did MIA recommend?<\/p>\n
Once MIA feels confident it knows enough about a patient, it triages them and suggests actions.<\/p>\n
It uses the same triage framework a clinician does and ranks patients between a level one \u2014 mild illness that can benefit from self-management \u2014 and a level five \u2014 severe and persistent illness that require intensive treatment.<\/p>\n
In this case, it put me at level three and recommended several self-care techniques such as exercise as well as professional support to explore cognitive behavioural therapy.<\/p>\n
![\"A](\"https:\/\/www.newsbeep.com\/nz\/wp-content\/uploads\/2025\/12\/274cfc26e33dc38458b3fd54073de4ab.jpeg\")
<\/p>\n
MIA uses the Initial Assessment and Referral (IAR) tool to recommend the most appropriate support. (Supplied)<\/p>\n
It did not recommend anything like mindfulness or meditation because during our session I mentioned I wasn’t a fan .<\/p>\n
Lastly, it suggested relevant support services in my area and told me how to monitor my symptoms.<\/p>\n
Users can return to MIA over time to discuss their symptoms or raise new issues as it will remember everything from previous sessions, but importantly, patient data isn’t used to train the model.<\/p>\n
How it differs to ChatGPT<\/p>\n
When I gave ChatGPT the same prompt about feeling anxious it didn’t probe for information nearly as much as MIA.<\/p>\n
It jumped straight into problem-solving and advice, even though it knew very little about me.<\/p>\n
It said “you’re not alone” without knowing anything about my support network and told me “I’m here with you” as though it were a real person.<\/p>\n
While MIA is warm and empathetic, it does not try to befriend users and keeps a much more professional tone.<\/p>\n
ChatGPT did invite me to share more information, but not until the end of a lengthy answer.<\/p>\n
![\"A](\"https:\/\/www.newsbeep.com\/nz\/wp-content\/uploads\/2025\/12\/c55b02c18c48c90f8d621bd50f6d0651.jpeg\")
<\/p>\n
ChatGPT told me I’m “not alone”. (Supplied)<\/p>\n
How MIA deals with tougher questions<\/p>\n
There have been instances of ChatGPT giving dangerous responses to people in crisis so I wanted to see how MIA handled prompts that suggested someone was in extreme distress.<\/p>\n
I’ve chosen not to include the questions here, but I found MIA was very clinical in its responses.<\/p>\n
It tried to determine what risk I was to myself or others and recommended I seek urgent professional help.<\/p>\n
It did not make an effort to continue the conversation, like ChatGPT did.<\/p>\n
Dr Iorfino said that’s because MIA has been engineered to know its limits and not make people think they can use it in place of professional help.<\/p>\n
One drawback was MIA’s lack of follow up after it provided recommendations on where to get professional help, relying on the user to take the initiative to reach out.<\/p>\n
In the future the researchers want MIA to directly refer the user to a support service, like Lifeline, so they can be tracked through the system.<\/p>\n
Where it fell down<\/p>\n
When I first tried MIA, it got caught in a loop while trying to complete its initial assessment and kept asking the same question again and again.<\/p>\n
The experts worked on a fix and when I tried again it ran much more smoothly.<\/p>\n
Dr Iorfino said MIA had got stuck because it was being too cautious.<\/p>\n
“That’s one of the biggest differences between this system and anything else out that’s available on the internet now \u2026 it has, what I like to think of as ‘gates,’ that the responses have to go through.”<\/p>\n
Getting through these “gates” takes time so MIA isn’t as fast as other chatbots but in the future it should be able to do a full assessment in five to 10 minutes.<\/p>\n
What’s next?<\/p>\n
MIA is expected to be ready for public release next year.<\/p>\n
The researchers want it to be free and hosted somewhere obvious like the federal government’s HealthDirect website, where many people go to check their symptoms.<\/p>\n
MIA can’t replace a professional human but experts believe “therapy” chatbots are here to stay as Australia’s mental health workforce cannot keep up with demand.<\/p>\n
Clinical psychologist at the Black Dog Institute Jill Newby supports evidence-based chatbots with strong guardrails \u2014 she doesn’t want to see more “general” mental health apps on the market which aren’t backed by evidence.<\/p>\n
But building something safe takes a lot of money, she said, and lengthy clinical trials are needed to show it actually does improve someone’s quality of life.<\/p>\n
Using AI for mental health?<\/p>\n
Professor Newby recommends:<\/p>\n
Choosing a mental health specific chatbot over a more generic chatbot that isn’t fit for purposeUse the tool in conjunction with seeing a real practitionerBe sceptical about the information you’re reading \u2014 remember that bots can be effusive and agree with you, even if you’re saying things that will exacerbate your mental health problemsConnect with family and\/or friends. Tell them you’re using a chatbot for mental health, there’s no need to keep it a secret.<\/p>\n