Why we need a body to oversee how science is used by governments
Science and Politics Ian Boyd Polity (2024)
Say what you really think, Sir Ian. Politics attracts people who are comfortable with “lying” and “manufacturing social truth” and who “do not distinguish between fantasy and reality”, Ian Boyd notes in his book Science and Politics. Moreover, many scientists are falling “for the wisdom of crowds”, “following the money” or, worse, indulging in the “evil” of normative research by designing studies to confirm their preferences. Borrowing former US president Barack Obama’s scathing comment about lawyers, the author labels many of those researchers as “highly credentialed, high-IQ morons”.
AI tools as science policy advisers? The potential and the pitfalls
In short, I loved this book. In Science and Politics, Boyd — an ecologist and former adviser to the UK government — weaves together anecdote and reason to unpick the “troubled marriage” between research and government. He offers insights into life inside the “policy factory” as a participant and, frequently, a frustrated observer. His narrative never shies away from robust language, yet it somehow retains an upbeat tenor throughout.
The relationship between science and politics is, Boyd laments, “too much theory and too little practice”. Boyd is very much a practitioner of science advice, having spent seven years (2012–19) as chief scientific adviser at the UK Department for Environment, Food and Rural Affairs (Defra). He also had a key role in developing science-advisory structures during the COVID-19 pandemic.
I overlapped with Boyd’s tenure at Defra while I was leading the Parliamentary Office of Science and Technology, the UK Parliament’s in-house source of research analysis. For years, I asked science and policy researchers how we could provide better advice, but the answers were rarely useful. This book, with its shamelessly practical bent, would have helped a lot.
There’s much in here, too, that tallies with my current role as a practice-focused academic who helps executives and legislatures to improve their use of evidence in decision making. Government clients usually ask me how they should design systems and processes. Although that is important, my starting point focuses on how to improve institutional cultures and the quality of staff.
How to spend one trillion dollars: the US decarbonization conundrum
Boyd agrees, and shares example after example of how, in the UK government, the truth is too often considered toxic and politicians and civil servants are unqualified or unable to hear it. For example, one senior civil servant told him how it was their job “to keep science advisers in their ‘boxes’”. Boyd recalls the “horror” on one UK cabinet minister’s face when he tried to explain something in theoretical terms.
While acknowledging flaws in the system, Boyd notes the benefits of the UK model of having chief scientific advisers in each ministry — individuals who are responsible for bringing research evidence to bear on policy. It won’t work when there is a bad adviser in the post, but a good one can be very effective at honing policy options, identifying unintended consequences and stopping bad ideas.
The rise of AI in policy
Looking to the future, Boyd flags how artificial intelligence (AI) promises to alter the functioning of government institutions. For example, ‘digital twins’ are increasingly used to test policy options. These are computer models that mimic complex systems associated with, for instance, agriculture, ecology or public health.
AI tools will also put evidence scoping and synthesis into the hands of mainstream policymakers, an advance that will surely be a headache for science advisers (see C. Tyler et al. Nature 622, 27–30; 2023). It is easy to imagine how political figures might use AI tools to produce ‘policy-based evidence’ rather than ‘evidence-based policy’.
Until now, advisers have been responsible for delivering or contextualizing scientific evidence and understanding its nuances. This isn’t always easy, or popular. For example, Boyd reflects on how he was politically exposed for having to explain weaknesses in opposing evidence syntheses on the role of badgers in the spread of bovine tuberculosis. In a room filled with the most senior politicians, civil servants and the president of the farming union, Boyd overruled approvals for two badger culls, which led to “an embarrassing climb-down by the politicians”.
Because AI tools make more-advanced modes of decision-making more attainable — such as real-time policy testing that produces feedback leading to better outcomes — advisers will need to increase their monitoring efforts to maintain the standards of people who use science but who are not trained in it.
How can scientific rigour be maintained? Boyd recommends establishing a science office, modelled on the UK Office for National Statistics or the UK Office for Budget Responsibility, that holds policymakers and science advisers to account in their use of research. This is not a new idea, but it is the most high-profile argument I have seen for it so far.
Such a statutory authority would enforce a system of internationally accepted standards, such as “follow the guidance of science” and “formally state your reasoning if you choose to depart from that guidance”. The latter part exists in several documents, including the UK ministerial code. Debate over upholding these standards will be lively, but resolvable.
Decision makers need constantly updated evidence synthesis
Yet Boyd feels that the current guidance lacks teeth. He professes to feeling “coercion to conform to the current political will” and feared a “dark space” where he could “control policies by tilting the ‘evidence’”. He argues that it would have been preferable, as a scientist working in policy, to be responsible to an external authority.
Serious thought would be required to ensure that a scientific authority does not seem to be a form of technocracy aimed at overturning democracy. Boyd suggests a renewed commitment to “open science” and “scientific freedoms” and calls for making the scientists of this body visible. In my view, such a body would also do well to adopt modes of social-scientific research that involve public participation.
After finishing Science and Politics, and despite its horror stories about the peddling of “hocus-pocus” by “charlatans”, I still felt upbeat. Boyd reflects that bringing science to bear on policymaking is important because there is societal value in “merging reality with aspiration”. This is an apt description of this book and a fabulous summary of Boyd’s career.