It feels like trust in our doctors and healthcare system is quietly slipping away, both here in Australia and over in the US. At a big meeting held by the Australian Ethical Health Alliance (AEHA) in May 2025, the message thundered through: we, as healthcare professionals, need to act fast. I was on a panel talking about the rising tide of bad information, and we all agreed that trust has always been the very lifeblood of medicine. But once that trust starts to drain away, everything becomes incredibly difficult – things like getting vaccinations, screening for diseases, making shared decisions with patients about their care, and even the simple act of planning their recovery and discharge, especially to aged care. It’s a bit like having a bank account where trust is your only currency; if you spend it carelessly, you’re left with nothing to buy the things that truly matter.
What’s really changed isn’t just that there’s more inaccurate information swirling around, but how people actually come to believe things now. Professor Michael Kidd, speaking at the symposium, highlighted the huge health trends we’re seeing, like the urgent need for better health and how we can accurately promote it. GPs, in particular, are finding themselves in a tricky spot. They’re seeing social media, “health” content generated by AI, and even conflicting messages from official bodies all creating a jumble of information. These new sources often compete with, and sometimes even overshadow, the clinical advice doctors offer. Patients are showing up with screenshots from WhatsApp, summaries from AI tools, and little stories they heard on TikTok, expecting us to treat these equally with our national health guidelines and decades of professional experience. When doctors try to gently correct them, these conversations can quickly turn into arguments, all thanks to this age of information overload.
Imagine this common scenario, brought up by Professor Stacy Carter at the symposium: a parent comes in, asking for their child to be let off routine immunizations. They’re armed with claims about vaccine side effects, all neatly put together by AI, and demand tests that just aren’t backed by science. This puts doctors in a real bind. On one hand, we’re bound by duty to protect public health, follow legal vaccine reporting requirements, and order tests responsibly. On the other hand, we also need to keep that crucial relationship with our patient, focusing on minimizing harm. So, simply saying, “That’s wrong,” while perhaps ethically true, often doesn’t work in practice. The AEHA discussion made it clear that this spread of bad information isn’t just a communication hiccup anymore; it’s a serious threat to public health. We can’t just throw facts at patients when we’re up against powerful AI models. Instead, our job is to guide and empower them, showing them how to check AI-generated information and find trustworthy sources.
Many speakers at the symposium connected this problem of misinformation to something called “epistemic injustice.” This is a well-known issue where certain groups – like Indigenous Australians or people from culturally and linguistically diverse backgrounds – are less likely to be believed. If these same communities are also being targeted by online misinformation, then just saying, “Trust me, I’m the expert” isn’t just ineffective; it can actually strengthen old feelings of distrust that have existed for a long time. The suggestion was to move away from a top-down approach, where the doctor simply tells and the patient listens, towards a partnership. This new way of thinking, where we say, “Let’s look at your source and my source together,” was seen as a much more positive and collaborative approach. It’s about building a bridge, not a wall.
Now, about AI: is it making things worse or can it actually help? The people at the symposium were crystal clear: AI is already a big part of the misinformation problem. It’s churning out health advice that sounds convincing but is actually wrong, digging up information that hasn’t been properly checked by other experts, and sometimes taking medical facts completely out of their proper context. As Professor Chris Bain pointed out, digital tools, when brought in without strong input from doctors, can “add to misinformation… especially when detached from expert health input.” But this isn’t a reason to throw AI out completely; it’s a reason to manage it well. Doctors need to be the ones guiding these systems, setting the rules, insisting that the data used to train AI is diverse, demanding that we can check how it works, and even suggesting having an ethicist involved. If we don’t, we’ll just end up automating our existing prejudices and spreading them even faster. This is where good technical management really matters. Australian health services already know how to check for privacy risks and manage cyber threats; now they need similar processes for AI-generated content and tools that patients use, all tied into our existing clinical governance framework. This means having accountable committees that include both doctors and patients, a clear process for updating AI models, and quick ways to fix things when we find misinformation in a tool, whether by retraining the machine or correcting the content itself. The real danger is misinformation that goes unnoticed, or even deliberate fake news created by AI.
Sometimes, just presenting the facts isn’t enough. One striking example talked about at the AEHA symposium was the lack of expected change after children died from measles overseas. You’d think such tragic events would lead to a surge in uptake of the MMR vaccine, but they didn’t. The sad truth is that communities often know the facts, but they either don’t trust the source, or they prioritize things like their personal identity, their community’s beliefs, or individual stories over cold, hard population data or decades of evidence that clearly show childhood vaccination programs save lives. The panel suggested we need to rely more on trusted community leaders, use storytelling in primary care, and properly support clinicians who are doing this emotionally draining work, often within tight 15-minute appointment slots. It’s about connecting with people on a human level, not just a scientific one.
A really strong message from the symposium was that for health advice to be effective, people need to believe the person giving it. Nina Roxburgh put it simply: we need to move beyond expecting only official institutions to be credible. Instead, we need to share that credibility with patients, especially those who have direct personal experience with what we’re talking about. In plain English, this means communicating in a way that acknowledges past trauma, offering interpreters, designing health materials together with the communities they’re for, and documenting things respectfully – even if the patient’s online source of information is limited or inaccurate. These might seem like small gestures, but when many people do them, they add up and slowly rebuild trust in our institutions.
The people at the AEHA symposium weren’t naive; they knew how hectic general practices, emergency departments, and outpatient clinics are. So, the actions they suggested were deliberately practical, asking clinicians to correct misinformation respectfully, use consistent messages, and tackle common rumors so they can be dealt with efficiently on a larger scale. This includes clearly naming misinformation without making fun of it, inviting patients to look through information sources together, and making sure clinicians aren’t left to come up with answers on the spot. We also need to flag recurring misinformation to health service leaders so it can be fixed at a system level, like on websites or community channels, rather than just in one-on-one consultations. And finally, clinicians should advocate for AI and digital tools in their services to have proper clinical oversight, not just tech sign-offs, specifically to monitor for misinformation. Above all, the symposium underscored that trust today isn’t just given; it’s earned through our actions. We have to show we’re transparent, admit when we’re unsure, and clearly explain how we make decisions for patients. Because if we don’t, that emptiness will be quickly filled by AI-generated “certainty,” even if it’s completely wrong.