> Academics and experts are losing the salience war. Most don't know they're in one. ![[Gemini_Generated_Image_jwlmlgjwlmlgjwlm.png]] For years I thought Australia had a know-do gap. I have written on this — it is the persistent space between what we know and what we actually do about it. I discussed this in my artice [[The Translation Problem]] But now I think we have something more fundamental going on: a salience gap. Not a failure to act on knowledge, but something closer to a systematic exclusion of the knowledge that would make the hardest questions unavoidable. Let me explain. ## Not Invited I have worked in robotics in Australia for twenty-five years. In that time, the community I was part of made genuine, world-class contributions — particularly to the mining and agriculture sectors, where Australian autonomous systems became world-leading. But for most of that period, I understood that robotics was a niche industry. We worked in the background. We did not attract media attention, and we did not expect to. That began to change around 2014, with the global conversation about the Fourth Industrial Revolution. It accelerated significantly after 2017, when a report by AlphaBeta — *The Automation Advantage* — projected a $2.2 trillion opportunity (in Australian dollars) for Australia from automation by 2030.[^1] The report noted that 50% fewer Australian firms were engaged in automation compared to leading countries, and called for bold policy to transition workers into the higher-value roles that automation would open up. For the first time, the structural question — what robotics and AI would do to the shape of work, and what Australia needed to do about it — was being asked publicly, with numbers attached. The pattern that followed is one I came to understand only gradually. In 2020, I presented at the AI Group's Industry 4.0 conference on digital skills and reimagining work in advanced manufacturing.[^2] The structural question was well established by then: automation and robotics would reshape which jobs existed, which skills were valuable, and which training was worth investing in. The community had done the work. The question was whether the policy process would engage with it. In 2022, the Australian robotics community published the second edition of the Robotics Roadmap for Australia — a 240-page national strategy drawing on 1,058 consultation participants from across industry, government and research.[^3] It documented Australia's robot population density at 75, against a world average of 113. It forecast which sectors automation would transform and on what timeline. It called explicitly for a national retraining and upskilling program, and for national curricula to be updated to reflect digital and technology skills. It identified, in detail, which jobs robots would take from humans — and hence what skills Australia would need in their place. That same year, the Albanese government convened the Jobs and Skills Summit.[^4] Its explicit purpose was to address jobs, skills and the future of work. Its issues paper flagged automation, digitalisation and structural change as central challenges. Not a single robotics or AI expert was invited. The community that had done the most rigorous forecasting work on exactly the question the Summit was convening to address — which technologies would reshape employment, which skills would be obsolete, which training was worth investing in — was excluded from the room entirely. The Summit produced recommendations on skills, training and migration without the input of the people who could tell you which skills automation would eliminate within a decade. ## The Shell Game The government responded to the Robotics Roadmap by commissioning a National Robotics Strategy.[^5] National workshops were held across Australia. In May 2024 the strategy was published — acknowledging that automation could add up to $600 billion per year to Australia's GDP by 2030, that Australia ranked 32nd globally in the adoption of industrial robots (a separate measure from the density figure above — count per 10,000 workers versus international rank), and committing explicitly to monitor workforce changes to align skills development with advancements in robotics and automation technologies. Then in 2025, the consultation continued. In June, the Treasurer opened what was framed as a Productivity Roundtable focused on the role of technologies — AI, robotics, automation — that might lift Australia's stagnant productivity growth. I made a submission arguing that productivity improves through innovation, innovation is constrained by adoption, and adoption depends on the effective diffusion of knowledge, currently hindered by inadequate documentation and the absence of accessible archives of past innovation projects.[^6] I also raised a structural question: as robotics and AI steadily reduce the need for human labour in productive industries — mining, agriculture, manufacturing — employment shifts toward human-centred, largely government-funded sectors: health, education, aged care, emergency response. If companies need fewer people and pay less tax, who funds the society that allows them to operate? By the time the event convened in August, it had become the Economic Reform Roundtable — three days, five pillars, spanning resilience, skills, competition, care, and an entire day on tax. I welcomed the broadening. But AI and innovation were allocated a single two-hour session on the afternoon of day two. The structural question — what productivity even means, and who funds the care economy, once robots produce and humans care — never surfaced in that session, or anywhere else across the three days. For a long time, I assumed this was scope drift. No one's fault in particular. Many of these issues are wicked problems — genuinely complex, with no single owner, no clean solution, and no natural constituency to keep them on the agenda. When a roundtable broadens from productivity to economic reform, it isn't necessarily because someone decided to bury the hard questions. It's because the hard questions are hard. They resist the formats that policy processes are built around: the bilateral negotiation, the costed proposal, the stakeholder with a clear ask. Robotics and AI's impact on work and taxation doesn't fit neatly into any of those. So it drifts — not out of malice, but because the system has no mechanism to hold it. Then I read about Vicebio. ## The Moment It Crystallised In July 2025, Sanofi announced it would acquire Vicebio for US$1.15 billion upfront, rising to US$1.6 billion (around AUD$2.4 billion) with milestones.[^7] Vicebio was a London-headquartered company, created by a European venture firm to commercialise a technology discovered at the University of Queensland: the Molecular Clamp, a vaccine platform with direct pandemic-preparedness implications, built over twelve years by Professors Keith Chappell, Daniel Watterson and Paul Young. The science was Australian. The company that captured its value was not. The pharmaceutical trade press covered it immediately. BioPharma Dive. PharmExec. Health Industry Hub. The European wires. UQ's own communications team did everything right. The Australian — the country's national broadsheet — did not feature it on the front page that day. I checked the front page of The Australian for 22 July 2025 myself.[^8] Barnaby Joyce and net zero. Mark Latham's vulgar messages to Julie Bishop. A super tax dispute. Transgender issues in the Greens. Gout Gout's Commonwealth Games ambitions. A pub for sale in North Sydney. Dating after 60. Weighted vests. Not a word on the front page about one of the largest commercialisations ever to flow from Australian university research. I genuinely could not fathom it. Not anger, exactly — more bewilderment. How does this not matter? But then a more uncomfortable thought followed. What if the editorial judgment was simply correct? What if Australians genuinely don't care about an AUD$2.4 billion vaccine deal — not because it was suppressed, but because innovation and research capability don't register as things that affect their lives? People feel real estate prices. They feel migration. They do not feel the absence of a sovereign vaccine platform. The harm from underinvesting in science is real, but it is diffuse, delayed, and abstract — and the attention economy has no mechanism for things that are diffuse, delayed, and abstract. That is not a media failure. It is a deeper cultural one. And it is the one that no amount of better communication fixes on its own. I just didn't have a word for it yet. ## Salience The word arrived from an unexpected direction — a podcast by Gary Stevenson, the economist and former Citibank trader behind Gary's Economics, called "Refugee Riots, Salience and Storytelling."[^9] Gary tells a story about a radio interview early in his media career, with Nick Ferrari on LBC. Gary wanted to talk about inequality — how COVID was widening the wealth gap, how it would lead to an economic crisis. Ferrari kept pulling the conversation back to a single narrow question: what energy provider should ordinary families use? Again and again. Every time Gary tried to broaden the conversation, Ferrari steered it back to energy providers. Gary's realisation reshaped how I understood everything I had been experiencing: > *Salience is not about winning the argument. It is about controlling what topics are discussed.* That was it. That was the word I had been looking for. The Productivity Roundtable dissolving into the Economic Reform Roundtable. The silence around Vicebio. None of these were failures of trust, merit, or evidence. We were not distrusted. We were not even being argued with. We were simply not salient — not present in the frame within which the conversation was being had. From that point, I started asking a different question. Not "is this true?" or "is this important?" — but: who is controlling the narrative here? What is the thing not being talked about? And why? ## The Trust Gap That question reframes a debate that has been running in science communication for years. Scientists are among the most trusted professions in the world, and yet misinformation is winning, research funding is being cut, public health is under attack. The conclusion often drawn is that this is not a trust problem — it is an influence problem. That's close. A landmark 2025 study in *Nature Human Behaviour* surveyed 71,922 people across 68 countries and found trust in scientists moderately high almost everywhere — a global average of 3.62 on a 5-point scale, with 78% of respondents perceiving scientists as qualified and 57% as honest.[^10] The "crisis of trust" narrative, the study concludes, is substantially overstated. And yet — trust sits high across the board, and evidence-based policy keeps losing. If trust were the bottleneck, it would translate into influence over what gets discussed, funded, and legislated. It doesn't. Trust sits there, broadly held, while the policy agenda is set by something else entirely. Trust measures how much people believe you when they encounter you. Salience is whether you are encountered at all — whether you are present, prominent, and activated in the mind at the moment a decision is being made. Credibility without salience is invisible. This distinction has a substantial academic literature behind it. McCombs and Shaw's agenda-setting theory, developed from 1972 onward, established that media doesn't tell people what to think — it tells them what to think *about*. The transfer of salience from media agenda to public agenda is one of the most replicated findings in political communication research.[^12] What I am describing is the failure of that transfer — or more precisely, its capture. When the transfer mechanism works against scientific and innovation communities, and in favour of those with the resources and strategy to engineer it, the result is a salience gap: issues that are objectively important but structurally invisible. I watched this misdiagnosis play out at my own former organisation. In March 2024, after Opposition Leader Peter Dutton attacked the credibility of CSIRO's GenCost report — the modelling underpinning the cost comparison between renewables and nuclear energy — CSIRO's Chief Executive, Dr Doug Hilton, responded with an open letter centred on trust: > "Maintaining trust requires scientists to act with integrity. Maintaining trust also requires our political leaders to resist the temptation to disparage science." It is a message Hilton has returned to often — in the Thomas Baker Oration, his CEDA address, interviews about CSIRO's role. Trust, social licence, integrity. All true, all important, and all addressed to the wrong battlefield. The GenCost dispute was never really about whether CSIRO's scientists had integrity. Nobody seriously believed they were lying. It was a fight over which numbers got repeated on talkback radio, which framing of the energy debate took hold, which report got cited, by whom and how often — a salience and framing battle over nuclear versus renewables. Hilton answered it with a defence of trust. The framing answered the wrong battle. CSIRO's subsequent troubles — the funding pressure, the political vulnerability, the sense of an institution constantly on the back foot — are, in part, the cost of fighting the last war while a different one was underway. To be clear: trust in science is critical. It is the foundation everything else rests on, and Hilton is right to defend it. But it is not sufficient. Without salience, no one cares. ## How Salience Is Built Consider Australia's nuclear power debate. Nuclear remains prohibited under federal and state law. No reactor could plausibly be operating before the mid-2040s. A federal parliamentary committee found it would be far more expensive than the renewables pathway Australia is already pursuing. And yet for roughly two years, nuclear power dominated the national energy conversation. They did not win the argument on the merits. They won a salience contest. A peer-reviewed discourse analysis examined nearly 600 articles published between January 2015 and June 2022.[^11] The ABC ran 40 pieces, the Sydney Morning Herald 62, The Age 25. The Australian ran 285, The Adelaide Advertiser 181. Two News Corp mastheads produced roughly four times the coverage of the ABC, SMH and Age combined, using recurring hooks — AUKUS, energy security, "we're already a nuclear nation" — to regenerate attention whenever the debate threatened to fade. The case for an accelerated renewables build-out, requiring no change to the law and available far sooner, never received anything like the same attention, despite being the option CSIRO's own modelling consistently favoured. This is a **communication architecture failure** — an institutional inability to translate genuine credibility into the currency the attention economy runs on. The nuclear campaign understood the game. CSIRO, defending its modelling on grounds of trust, largely did not. Salience can be created by amplifying an issue. It can also be created by ignoring one. Vicebio illustrates the latter. Nuclear power illustrates the former. In both cases, the question is not whether the issue is true or important, but whether it occupies enough space in the public conversation to shape decisions. ## The Call I have had the privilege of talking to many leaders — in science, in politics — over the course of this work. Most are smart, capable, well-intentioned people. But their choices are constrained by what they can do, not by what they want to do. It is salience that gives them permission. A minister who privately agrees Australia under-invests in research still needs the political cover to act. A department secretary who knows the data is compromised still has to operate inside a public conversation that hasn't caught up. Salience is not a substitute for good policy. It is the precondition that makes good policy possible. So what does an institution that accepts this actually do? First, it stops fighting the trust battle when the fight is about salience. When CSIRO's GenCost modelling is attacked, the answer is not another defence of scientific integrity. It is to make the numbers travel — to get them repeated, framed and cited faster and more often than the alternative. The objective is not to be believed. It is to be present in the frame when the decision is made. Second, it builds salience-ready material as a matter of course, not as an afterthought. The evidence on energy costs, vaccine safety, and the return from public research is full of salience-ready content: dollars saved, lives protected, named populations benefited. It exists. It is simply never packaged in a form the attention economy can carry. The nuclear campaign understood the format — opinion pages, recurring hooks, sympathetic framing devices repeated until they stuck. The institutions sitting on the actual evidence still default to the press release and the journal abstract. Third, it treats the permanent record as infrastructure, not exhaust. The reason the Vicebio deal could vanish from the national conversation in a single news cycle is that there was nowhere durable for it to land — no accessible, linkable, provenance-backed account a journalist, a parent or a model could find six months later. Salience in the moment is necessary but perishable. What survives the news cycle is what has been deposited somewhere findable. A country that does not maintain that record is choosing, by default, to be narrated by whoever does. And salience begins long before policy is made. It begins around dinner tables, in schools, and in the stories a country tells itself about what matters. The most important people to influence are not ministers or industry leaders. They are parents. If parents believe science and innovation matter to their children's future, politicians follow and industry follows. If they do not, neither will anyone else. ### A Personal Note I have spent the past ten years trying to raise awareness of robotics in Australia. Presenting to industry and the public. Writing submissions to government. Contributing to three editions of the Robotics Roadmap. All of it leading, eventually, to a National Robotics Strategy — the first in the country's history, published in May 2024, promising up to $600 billion per year to Australia's GDP by 2030. Then the government lost interest. No implementation funding was announced. No progress report has been released. The strategy sits on a website, unreferenced by the ministry responsible for it. We lost salience. At the time, I posted a link to Moving Pictures / Shannon Nole - *What About Me*.[^13] This is a personal issue for me, because I came to understand it late in my own career. Toward the end, I started doing outreach — talking to students, going into schools and universities. I told myself the point was encouraging the next generation into STEM. It wasn't, really. Looking back, I am not an educator. What I was actually doing — without having the word for it — was trying to make our work *salient*. Putting it in front of people, in a form they would notice and remember, before the moment came when it might matter. Not recruitment. Presence. The same challenge applies to every institution that holds Australian expertise, and to those of us who have spent a career producing knowledge and assumed that producing it was enough. Academics and experts are losing the salience war. Most don't know they're in one. The nuclear campaigners, the lobbies that flood talkback radio, the media operations that manufacture outrage — they all understood something we didn't: being right is not the same as being present, and presence takes deliberate, sustained, often uncomfortable effort. But salience is only the first layer. Once attention is secured, a second battle determines how issues are framed. A third determines what evidence is available when those issues are debated. The three reinforce one another. This article is about the first. A second article — *We Are Losing the Data War* — takes up framing and the evidence that underpins it. ## References [^1]: AlphaBeta, [*The Automation Advantage:](. https://accesspartnership.com/wp-content/uploads/2023/03/The-Automation-Advantage.pdf) How Australia can seize a $2 trillion opportunity from automation and create millions of safer, more meaningful and more valuable jobs* (2017), [^2]: AI Group Industry 4.0 Advanced Manufacturing Forum, ["Digital Skills for Industry 4.0 – Reimagining Work in Advanced Manufacturing](https://i4amf.aigroup.com.au/i4amf-conference-2020-digital-skills-for-industry-4-0-reimagining-work-in-advanced-manufacturing/)" (2020). [^3]: Keay, S. et al., [*Robotics Roadmap for Australia 2022*](https://bluezonegroup.com.au/wp-content/uploads/2022/01/Robotics-Roadmap-for-Australia-2022_compressed-1.pdf), Robotics Australia Group (2022). [^4]: Australian Government Treasury, [*Jobs and Skills Summit Issues Paper* ]( https://treasury.gov.au/publication/p2022-307396)(17 August 2022). [^5]: Australian Government Department of Industry, Science and Resources, [*National Robotics Strategy*](https://www.industry.gov.au/publications/national-robotics-strategy) (May 2024). [^6]: Duff, E. [[2025-08 Submission to Commonwealth Treasury on Economic Reform Roundtable]] [^7]: University of Queensland, "[Billion-dollar deal takes UQ vaccine tech to the world]( https://news.uq.edu.au/2025-07-billion-dollar-deal-takes-uq-vaccine-tech-world)" (July 2025). [^8]: The Australian, [front page archive,](https://archive.is/IdHjx) 22 July 2025. [^9]: Gary Stevenson, "[Refugee Riots, Salience and Storytelling,]( https://www.youtube.com/watch?v=9F4OSDONAR4)" Gary's Economics (YouTube, 2023). [^10]: Cologna, V., Mede, N.G. et al., "[Trust in scientists and their role in society across 68 countries,]( https://doi.org/10.1038/s41562-024-02090-5)" *Nature Human Behaviour* (2025). [^11]: Altenkamp, I. and McManus, P., "Nuclear power in a de-carbonised future? A critical discourse analysis of nuclear energy debates and media framing in Australia," *Australian Geographer* (2024). [^12]: McCombs, M. and Shaw, D., "The Agenda-Setting Function of Mass Media," *Public Opinion Quarterly* (1972). For a recent application to science policy, see Bromley-Trujillo, R. and Karch, A., "Salience, Scientific Uncertainty, and the Agenda-Setting Power of Science," *Policy Studies Journal* (2021). [^13]: Duff, E., LinkedIn post, May 2024 [[2024-05-04 What About Me]]