A jab well done: How Australia’s vaccine strategy can fight misinformation

Australia’s childhood vaccination coverage has fallen for five consecutive years. Over roughly the same period, the information environment shaping parental and community decision making on vaccine uptake has been fundamentally transformed by generative AI. That doesn’t make AI the sole cause of declining coverage. The effects of COVID-19, rising institutional distrust, global political polarisation, and renewed anti-vaccine activism all matter. But it does mean Australia is now trying to rebuild vaccine confidence in an information environment that no longer behaves like the one to which public health officials have long been accustomed.

The latest national data make the challenge unambiguous. The National Centre for Immunisation Research and Surveillance (NCIRS) reported in April 2026 that full childhood vaccination coverage fell in 2025 to 90.5 per cent at 12 months, 88.4 per cent at 24 months, and 92.5 per cent at 60 months. Around 80,000 Australian children were not fully vaccinated at those three early-age milestones. Adolescent HPV coverage also remains well short of the 90 per cent target needed to support cervical cancer elimination, sitting at 78.7 per cent for girls and 75.6 per cent for boys.

The equity picture is similarly striking. NCIRS highlighted marked decreases in vaccination coverage and on-time vaccination among Aboriginal and Torres Strait Islander children and adolescents, reinforcing the need for culturally safe, community-led responses. Coverage decline is also geographically uneven: vaccination rates in the lowest-coverage areas are falling much faster than in the highest-coverage ones. In practical terms, communities already facing access barriers, lower institutional trust, or weaker health-service reach are being asked to absorb a larger share of the risk.

This is now a public health problem in Australia. Coverage declines are epidemiologically meaningful because they may erode herd protection, increasing the risk of population-level transmission of vaccine-preventable diseases (VPD), not just individual acquisition. Australia is not alone in this downward trend – VPDs are resurging globally. The costs are borne by children, families, communities, clinicians, and health systems. It meanwhile becomes harder to rebuild trust after every missed opportunity to prevent avoidable harm.

Australia has not failed to diagnose the problem. The COVID-19 Response Inquiry recommended a national strategy to rebuild trust. The National Immunisation Strategy 2025–2030 names misinformation and disinformation as first-order threats to community understanding and acceptance. An Australian survey of parents published by Kaufman et al identifies a decline in parental confidence in vaccination from 2024 to 2025, with growing exposure to misinformation – particularly from the US – an important contributor.

The Australian Communications and Media Authority defines misinformation as false, misleading or deceptive information that can cause harm. Disinformation is misinformation that is spread with the specific intent to deceive. Our focus here is on misinformation, but the issue is broad. Vaccine misinformation obviously predates generative AI, which shouldn’t be allowed to become the go-to explanation for every fluctuation in public trust. The COVID-19 pandemic disrupted routine care, politicised public health, and saw unprecedented attention targeted towards public health emergency responses by the public, including on vaccine development, evidence and effectiveness. International politics, including high-profile anti-vaccine rhetoric in the United States, has further legitimised scepticism that once sat closer to the fringe. In the midst of this scrutiny, AI has changed the operating conditions by which vaccine misinformation now spreads.

Generative AI represents a significant change in the scale and effectiveness of misinformation through the combination of production capacity, apparent credibility, personalisation, and reach. A January 2026 BMC Public Health systematic review authored by Hamid Reza Saeidnia and colleagues found that generative AI lowers barriers to producing convincing false health stories, increases the speed and perceived credibility of disinformation, and exposes weaknesses in existing detection and labelling systems. Research conducted by Nicholas B. Tiller and colleagues in April 2026 in BMJ Open audited the leading chatbots and found that nearly half of their responses to health and medical misinformation prompts were problematic, with hallucinated or unreliable citations compounding the credibility problem.

The format of misinformation has also changed. It is no longer confined to amateur blogs, poorly designed websites, or fringe Facebook posts. Health misinformation now moves through monetised podcasts, short-form video, influencer ecosystems, and identity narratives. Deepfake clinicians and synthetic experts can make false claims feel local, authoritative, and emotionally resonant. An analysis by AAP FactCheck in 2025 found a video claiming to show Canada’s chief public health officer warning about dangerous COVID-19 vaccine side effects was an AI deepfake, with several versions distributed across social media platforms to promote a spurious blood-clotting product.

Examples like this define the AI-era challenge. The public health system now faces misinformation rooted in narratives of identity and empowerment that make conventional corrections psychologically ineffective.

Australia has made important strategic and institutional advances. The Australian Centre for Disease Control, established since January 2026 as a permanent national body, identifies health misinformation as an immediate priority. The National Immunisation Strategy names six priority areas, including rebuilding trust, understanding, and acceptance, and harnessing new technologies to respond to the changing vaccine and communicable disease landscape. Its Implementation Plan creates a clearer Australian Government and state role architecture and identifies practical initiatives such as SKAI (Sharing Knowledge About Immunisation) resources for clinicians, PHN immunisation support, First Nations uptake support, catch-up tools, and improved reporting. Clearly government has neither ignored the issue nor failed to develop a toolkit to tackle it. The problem is that the toolkit still looks substantially like that we were using prior to 2020: clinician training, conversation guides, community engagement, coverage dashboards, and public communications.

Those tools remain necessary, but they are no longer sufficient on their own. Australia doesn’t yet have a dedicated real-time surveillance function for health misinformation, an AI-augmented rapid content response capability, a standing deepfake takedown protocol for impersonated Australian clinicians, or a health-specific connection to provenance, watermarking, and platform standards work.

The gap as we see it is the operational capability level – technical, funded, and sustained – that would let the National Immunisation Strategy deliver on its own commitments at AI-era scale, with AI forming part of the solution.

We see four capabilities as key to matching the scale of the problem to the solution, with AI as a critical enabler to the response. This approach is consistent with the National Immunisation Strategy priorities, specifically Priority 2: Build trust, understanding and acceptance of immunisation in communities, and Priority 5: Harness new technologies to respond to the evolving communicable disease and vaccine landscape.

AI-enabled intelligence. Australia needs continuous technical surveillance of the health misinformation environment. Funded properly, this would look like a health-specific misinformation observatory housed within the Australian CDC, using AI-assisted content monitoring, community listening, and behavioural analytics to identify emerging narratives, affected geographies/communities, websites and other digital platforms used to spread misinformation, and the questions parents and others are actually asking. It should publish regular public assessments and feed clinician-facing tools such as SKAI in real time.

Rapid content response leveraging AI. The response to industrialised misinformation cannot rely on content pipelines that move in weeks when false claims move in hours. A funded capability would combine clinical evidence, behavioural science, and responsible AI tools to produce segmented, community-appropriate responses quickly. The same personalisation that makes misinformation powerful can be deployed ethically to make credible information more relevant, timely and persuasive. Accessibility of this content to generative AI systems will also enhance discoverability, and ultimately accuracy, of the general-purpose chat tools, which are increasingly becoming the first port of call for many of us.  

Platform and technical accountability. Even without a current misinformation and disinformation bill, meaningful levers remain. Governments can use procurement standards, provenance expectations, watermarking commitments from AI providers, health-specific transparency obligations, and co-designed takedown protocols for deepfake impersonation of Australian clinicians. Health can’t be allowed to sit outside the broader AI-safety and platform-regulation agenda. It needs dedicated representation within it.

Coordination and governance. These capabilities cut across the Australian Government, the Australian CDC, state and territory health departments, PHNs, professional colleges, and research centres, including NCIRS and MCRI. The Australian CDC is the natural coordinating home for the response, but only if it is given a clear mandate, standing funding, and specialist staff in data science, digital analytics, behavioural science, and platform engagement.

Industrial-scale, high-quality misinformation is now a permanent feature of the health environment. More than how they might win the communications battle, the strategic question for Australian health policymakers is how to build durable technical capability, funded across cycles, staffed with the right expertise, and coordinated across levels of government.

The National Immunisation Strategy and its Implementation Plan set the direction. The coming Commonwealth and state health budgets are where the strategy either becomes an AI-era response or remains a pre-AI response scaled up.

Get in touch to discuss how your organisation can build the technical capability to turn the National Immunisation Strategy into an AI-era response to vaccine misinformation.

Connect with Geoff Sjollema on LinkedIn. This piece was co-written by Stephen Philpot during his time at Nous.