What a biotech developer’s question reveals about social licence, science communication, and democratic accountability

A colleague of mine recently attended a meeting – one of those gatherings designed to bring different voices together on the future of agricultural biotechnology. During a break, a developer approached her with a question. Not: what are your concerns? Not: what is your evidence? Not: what would good governance of this technology look like?

Instead, his question was: what do I need to say to make you feel more positive about genetically engineered crops and foods?

She was taken aback – less by the sentiment, which anyone who has spent time in these spaces will recognise, than by the setting. This was supposed to be open dialogue. And here was someone cutting straight to the point: tell me the magic words that will make you love me.

The developer wasn’t concealing an agenda. He simply couldn’t see that “what do I need to say?” and “what are your concerns?” are entirely different questions. One treats the other person as a target to be won over; the other treats them as someone with knowledge worth hearing. That distinction sits at the heart of what is wrong with how the agricultural biotechnology sector talks about social licence.

Social licence as a communications problem

The term social licence may have the warm fuzzy kumbaya of modern dialogues, but it comes from the extractive industries – mining, oil and gas – where it emerged in the 1990s as a way of managing community opposition to projects already decided. Its logic was managerial rather than democratic: how do you secure enough acceptance from affected communities to proceed without costly disruption? Whether to proceed or not was never on the table.

When the concept migrated into biotechnology it brought this logic with it. In industry and regulatory circles, social licence has come to mean a measure of public approval – a threshold of positive sentiment that, once reached, is taken to justify a technology or product. You measure it through polling, focus groups and citizen dialogues. You achieve it through a communications strategy. If you don’t have it, you need better messaging.

This becomes very clear when polling produces inconvenient results. The UK’s Food Standards Agency has repeatedly surveyed public attitudes to genetic engineering (GE) in food, including gene editing and “precision breeding”. Consistently, those surveys (see here, here and here) show that people’s attitudes are at best conditional and at worst they just don’t want it. The response has not been to reconsider the policy. Instead, it concludes that people need to be better informed – that the gap between public sentiment and policy is a communications failure, not a democratic signal.

This is consent engineering: the biotech developer’s question, asked by government. The investment has been made, the endpoint is fixed and the technology will be introduced. The task of public engagement is to bring opinion into line with that endpoint, not to let opinion influence it.

The academic literature has a name for the assumption underneath: the deficit model – the idea that public scepticism about science and technology is always and only a product of insufficient information. It has been critiqued and discredited for decades, and yet it persists.

Research consistently shows that public concern about gene-edited food is driven less by ignorance than by values – about power, about who bears the risks, about ecological relationships, about who controls the food system and for whose benefit. People who understand the technology perfectly well can have coherent reasons to object; people who don’t understand it also have a right to reject it on their own terms.

The deficit model persists not because it is intellectually credible but because it is institutionally convenient. If public concern is always a knowledge problem, you never have to negotiate with it. You only have to find the right narrative.

The upstream problem: engineering consensus

Consent engineering, however, is just the visible part of the mechanism. Before the communications work begins, before consultations are convened and public approval is measured, the space for debate has already been shut down through the manufacture of consensus.

In settled science, consensus is an outcome – the product of accumulated independent evidence and of researchers converging from different starting points. The consensus on human-caused climate change was won this way, through genuine scientific contest, and has survived sustained attempts to disprove it.

The consensus claimed in food biotechnology looks different – not least because it leaves out the scientists who don’t buy into it. Four features stand out:

Repetition as evidence. Claims about the safety and benefit ofGE crops and foods have been repeated so often – in official documents, industry communications, regulatory frameworks and sympathetic science journalism – that repetition has become a substitute for evidence.

The statement “GE crops are safe” moves from a contested scientific proposition – contested not by cranks but by ecologists, toxicologists, epidemiologists and independent food scientists – to an assumed starting point. It acquires the social weight of consensus without the process that produces it.

Dismissal of dissent and refusal. In normal science, dissent is engaged with on its merits. In agricultural biotech, scientists who raise questions about gene flow, microbiome effects, toxicity and allergenicity, long-term ecological impacts or the adequacy of testing are instead labelled – as anti-science, anti-GMO, ideologically motivated or not credible. This is not how scientific disagreement works; it is how manufactured consensus protects itself.

Agreement that isn’t independent. Similar conclusions across regulatory bodies, science academies and research institutions can look like independent confirmation – the hallmark of real scientific consensus. But they can also reflect shared cultures, funding, personnel and ideological commitments.

In the UK, the precision breeding deregulation process offered a clear illustration: the same framing, the same conclusions, sometimes the same individuals on expert panels convened by different government departments – and the same language in Defra press releases, guidance documents and industry commentary. That tells you something about who is talking to whom. It tells you rather less about the underlying science.

Blurring of different kinds of questions. This is, arguably, the most consequential. A narrow technical question, such as “Does this particular GE crop produce a protein that is directly toxic in short-term animal feeding trials?”, is addressed in tests commissioned by the developer. Any adverse effects are argued away or downplayed, and the reassuring conclusion of “no risk” is routinely accepted by government advisors and regulators. That assertion of safety is then generalised through a series of leaps.

These leaps are fallacies that harden into fact: “Currently approved GE crops, in the opinion of some industry-aligned scientists and regulators, have not been shown to cause direct harm in short-term trials” becomes “GE food is safe” becomes “there is scientific consensus on GE food safety” becomes “opposition to GE food is anti-science.”

The final statement does enormous political work, but it rests (ironically) on a process of strategic editing that strips an already narrow claim of its qualifications. Consensus engineering is largely the management of those ‘edits’ – a quick, near-invisible sleight of hand.

The closed loop

What makes this system so resistant to challenge is that the two forms of engineering reinforce each other.

Manufactured consensus does the upstream work of defining what counts as knowledge, which questions are open and which objections are legitimate. Once “consensus” has answered the safety question in advance, safety objections are reclassified, automatically, as ignorance or ideology. The infamous Overton window has been set before public engagement begins.

Consent engineering then operates within the window the claimed consensus has created. The communications task becomes: how do we convey our message to people who haven’t received it yet?

This is why the biotech developer’s question felt so natural to him. He already knew the correct answer; he just needed the right way to deliver it. That my colleague might bring knowledge, reasoning or values the consensus hadn’t accommodated simply didn’t compute. It had been ruled out before the meeting began.

The loop is closed. Engagement creates the appearance of dialogue while the destination stays fixed, and anything short of acceptance is recycled as evidence of a communications problem. No citizen, farmer, ecologist, ethicist or civil society organisation can say “we have considered this carefully and we don’t want it” and have that count. The only recognised outcomes are acceptance and not-yet-acceptance. Refusal is not an option.

Breaking the loop

The closed loop is not inevitable. But breaking it requires more than better-designed consultations. It requires deliberative processes that allow citizens to reach conclusions the consensus has not already approved – and institutions willing to take those conclusions seriously.

There are recent examples from this very sector. One is the CARMA Citizen Forum, whose first-year findings were published in the January 2026 report Cultivating Debate. Eighteen members of the public – recruited for demographic diversity, not sympathy for the subject – spent more than twenty hours over a year deliberating on cultivated meat with scientists, regulators and critics. They were treated as citizens rather than consumers, and CARMA scientists were expected to respond to their guidance.

The Forum did not deliver a verdict for or against cultivated meat. Instead it set out political and economic conditions – not just safety thresholds – for it to contribute to a fairer food future. These included a non-commercial governing body with real oversight powers, strict limits on concentration of intellectual property, a rule that no single company could own more than one part of the supply chain, time-limited production licences, and long-term independent health monitoring. These are not the outputs of a hostile consultation but of thoughtful citizens given time to reason carefully – and they ask far more than the sector’s current business model could readily accommodate.

The Nuffield Council on Bioethics, with BBSRC and Sciencewise, ran a similar exercise in 2022 on genome editing in farmed animals. Eighty participants deliberated over four workshops, and what emerged confounded the deficit model at almost every turn: the more they understood about the technology, the more precise their concerns became – not the more accepting.

Crucially, participants saw the framing of genome editing as “like conventional breeding” as a device that obscured the questions they believed needed debating. They concluded that a genetically altered animal was a new kind of entity, and that the framing was being used to close down questions that should remain open.

They also pushed beyond safety – insisting that alternatives to genetic engineering be considered on equal terms, and that public aspirations rather than market forces should shape the research agenda. They wanted to be treated as citizens, not as consumers who would eventually accept what had already been decided.

A third example comes from closer to home. A Bigger Conversation’s Agroecological Intelligence report found much the same among forty-eight agroecological farmers and growers. They were not anti-technology; many already used apps, data analysis and precision tools. What they objected to was the narrative that technology is the main answer to sustainability challenges – a narrative shaped, as they accurately saw, by think tanks, developers and entrepreneurs whose interests are not those of farming communities or ecological health.

All three processes were designed – imperfectly, but meaningfully – to take seriously what participants brought rather than to correct it. And in each case, people kept arriving at questions the manufactured consensus had declared answered: who controls this technology, and for whose benefit? What are the alternatives, and why aren’t they being considered? Is this actually a different thing from what it’s being called?

These are not the questions of people who have not understood. They are the questions of people who have understood rather too well.

The question we actually need to ask

The honest answer to the biotech developer is: There is nothing you can say, but there are things you need to do.

Genuine social licence, if the term is to mean anything beyond a communications target, would require a different question altogether: what would it take for your concerns to count in how this technology is used and governed?

That question has practical consequences. It means oversight bodies with real power, as the CARMA citizens proposed, not advisory functions that can be quietly sidelined. It means accepting that deliberation might produce verdicts – not just conditions – that constrain or redirect technological development. It means treating consistent public rejection as a signal to heed rather than a problem to fix. And when deliberation produces inconvenient conclusions, it means asking whether the response is engagement or management.

Most fundamentally, it means abandoning the premise that the technology’s introduction is a fixed point, and accepting that in a democratic process, whether to use the technology at all is on the table.

The deliberative processes described here show that the window can be kept open, and that when it is, citizens ask better, tougher questions than the industry would prefer. Whether those questions are ever allowed to produce different answers is, ultimately, a question about democratic governance.

It is a question that the food biotechnology sector, and the regulators who support it, have yet to seriously face.

  • Pat Thomas is the co-founder and director of A Bigger Conversation.