I am about to complete my third and final three-year term as science advisor to the Prime Minister of New Zealand, having served three prime ministers across two administrations. While it is tempting to be reflective, I would rather focus this last essay on some of the ongoing and emerging issues that confront the science-policy and science-societal interfaces across the globe.
Digitalisation, broadly defined to include technologies such as artificial intelligence (AI), machine learning and the internet of things (IOT), brings with it many opportunities, but the challenges are far broader that just thinking about the future of work, or its impact on our goal of more inclusive societies or on the rise of the so-called ‘post-trust / post-truth’ environment. Clearly there are broader issues of privacy, autonomy and agency that are challenged by the power of these technologies. This technological evolution is also impacting on implied and essential contract between the citizen and the State: this contract could be further undermined in the future, if we do not sustain better and more deliberative conversations between industry, policy and science and innovation communities.
Over the past few months, through INGSA I have been involved in exploring the issues of human wellbeing at the face of digital transformation. The impetus for this came from a request from the OECD following discussions at the advisory group for the Going Digital project of which I am a member. We have worked on this issue over several months with some very engaging inputs, including from an expert workshop in London a few weeks ago and our report is now close to completion. It will highlight areas where more conversation, research and policy development is needed. Key to making progress is to consider a broad definition of wellbeing. It is not just based on self-perceptions of mental and physical health, nor is it simply about the material conditions that can influence this. Wellbeing is also about interpersonal, group and societal dimensions. In respect of the latter, if the relationship between the individual and the State is compromised by a decline in trust, then wellbeing – broadly defined – is affected.
The ongoing discussions about the future of work are much more than simply about automation and artificial intelligence. These changes have implications on individual wellbeing and indeed on the wellbeing of the State as it may affect tax bases significantly. The sense of social inclusion may be either increased or undermined by many factors, but digitalisation certainly will play a role in what emerges. Personal security is already affected by hacking and cyberthreats on one hand and yet greater possibilities for public safety surveillance on the other. How will the balance between technological possibilities, human perceptions, and human and social development play out? How do we promote greater resilience to change? How will we ensure the protection and reinforcement of critical human values like empathy? How will we educate people into the future? These are critical and poorly researched areas of enquiry. I think their urgency is underestimated, especially the need to focus on early human development and education to ensure greater wellbeing into the future.
In parallel, the Joint Research Centre of the European Commission has started a project entitled Enlightenment 2.0 which is looking at the question of how the relationship between information, policy making and society is being affected by this and other social transformations and in turn affecting political processes. These are important questions that require many disciplines from the humanities, social and natural sciences to work more closely together.
The issues of social acceptance and rejection of new technologies, whether digital or from the natural and life sciences, will continue to challenge us. Today, a broad range of technologies ranging from artificial intelligence, brain-machine interfaces, geoengineering or meiotic gene drive to gene editing may confront societies and policy-makers, to say nothing of the technologies that are yet to come. The possibilities are exhilarating but the unknowns are concerning. This tension requires a less polemical and much more reflective set of conversations about the meaning of precaution. Humans have always lived in experimental societies and no innovation is possible without some level of uncertainty, but is there agreement on how a society can come to collective understandings on such issues?
The ability to collect, analyse and use data is progressing rapidly. It is fundamentally changing ways in which knowledge is created and applied. With this, the Baconian model has in some cases been reversed and much science is now being undertaken in an almost hypothesis-free manner. And while the classical experimental method of science has often been more an ideal than a reality, data without interpretation and expert modelling can be misleading. The ‘association is not causation’ cliche becomes even more important to remember.
The modelling and computation approaches we now have allow researchers to ask questions of complex non-linear systems that up to now were beyond our abilities. Progress in areas as diverse as cosmology and climate change have benefited enormously from the extraordinary progress that has been made in computation and data analytics and which quantum computing, AI and machine learning will no doubt accelerate. The potential for many sciences to make great insights using these technologies is obvious.
We are starting to see this power being applied into the social sciences. The SESHAT anthropological and historical database is one compelling example. My particular interest has been in how data can assist better policy making in areas of social policy. Traditionally governments have seen health, education, social development, housing, justice etc largely through the distinctive administrative silos of separate agencies. But for the individual citizen, all these dimensions are absolutely intertwined. The challenge facing most democracies is how to better integrate across these and other dimensions of policy making to better fulfill government responsibilities. In this well curated and appropriately interpreted and analysed big data can lead to important new insights that can help.
But there is much more to the use of data than collecting it, cleaning it and analysing it. There needs to be social license to do so and that in turn requires ongoing and better public conversations, engaging both political and scientific leaderships. Citizens and societal groups need to know how data are going to be used for their benefit and what happens if something goes wrong. In my 2017 essay on Citizen-based analytics, I discussed some of these issues. There is no doubt that data can be of immense value in making better policy decisions but the social license to do so has to be developed and maintained. Data in isolation from context and from interpretive expertise can be misread, and thus the potential for AI to both be helpful, but also to be problematic is very high. Robust, transparent, inclusive and accountable data governance regimes are essential. Data governance and proactive oversight remains a marginal priority for many governments and this needs to be addressed if data is to play its potential role in achieving many goals that citizens expect.
As the impact of the Anthropocene becomes clearer, the importance of collective as well as individual action becomes clearer also. Yet there is always a tension between self-interest and collective interest whether we are talking about individuals or nation states. The slow progress on climate change reflects this. The problem defined by climate science is clear. The solution requires complex tradeoffs in which different interests compete, whether these are between the global north and south, across generations or between sectors. Many technologies may have a role in finding solutions but persistent and complex issues of societal acceptance, of perceptions and realities of who benefits and who bears the burden, of risk and precaution will likely continue to confound progress. Societally engaged science has much to contribute in addressing these issues.
The sustainable development goals (SDGs), which the global political community endorsed in 2015, encapsulate the manifest need to address these major issues of equity, equality, inclusiveness, economic and environmental sustainability. All countries signed up to these and they were never intended simply to be applied in developing countries. Yet there is a danger that they will have little or no pull on policy in many countries and be used simply as a reporting framework for business-as-usual decisions. INGSA and ICSU (soon to become ISC) have been working together to develop an approach that can help demonstrate the essential value of the SDGs on domestic policy processes. The approach focuses more on interactions between goals rather than on singular goals themselves. Hopefully some pilot implementation studies can be conducted in the next year. INGSA will also release its policy manifesto focused on the role of science advice in relation to the SDGs at its biennial meeting in Tokyo in November 2018.
Many have written about the so-called ‘crisis in science’. Most of the focus has come on the issue of reproducibility and faulty use of statistics. Some of the focus has been on the integrity of science and scientists. Others have focused on the influence of questionable publication practices. But there are deeper issues. The changing culture within science and within the institutions of science associated with its massive expansion have in one sense, ‘industrialised’ science. This is not a reference to the relationship between science and the private sector but rather to the way in which science is undertaken which now responds more to external incentives and drivers. These include impact factors of publications, institutional rankings and individualistic incentives that can undermine the promotion of interdisciplinary and team based research of high quality.
Science is an institution and in the interconnected world, trust in institutions is fragile and science must continue to evolve in ways that sustain the public’s trust. In this, we are making progress. Increasingly the scientific community is now better at holding the mirror up to itself, assisted by the work of critical social science scholars and asking difficult questions about our own institutions. But funding bodies, academies and research institutions will need to continue to work together to find ways to strengthen the institution of science while ensuring its inclusivity and engagement.
As science becomes necessarily more open, both in data and in its relationship to the societies in which it is embedded, we can and must maintain our rigour. Sadly, there are some who would capitalise on the imperative for openness; the rise of predatory journals and conferences is a symptom of the challenges we face.
These issues have long been the focus of science studies and science policy scholars and their arguments have become ever more relevant. Michael Gibbons wrote persuasively about this in his seminal article in Nature in December 1999. Ravetz, Funtowicz, Jasanoff, Weingard and many others in the STS tradition have produced a significant literature and insights that are less known within the conventional disciplines, but that are more relevant than ever. Concepts of co-production, co-design and extended peer review will increasingly be brought into practice.
I do not particularly like the ‘post-trust’ or ‘post-fact’ descriptors, but they are powerful in describing the challenge we all face in separating robust and reliably developed knowledge from rhetoric. The accessibility of networked knowledge has been a great advance for society and for science, but the confusion of unreliable or intentionally false information with robustly derived knowledge has clearly grown. Nor should scientists overstate their claims; there are other sources of knowledge that have validity to those who have access to it. Scientific hubris only reduces trust in science and limits its ability to have optimal impact.
The knowledge disciplines including both the social and natural sciences can both address our innate thirst for discovery and inform solutions to the multitude of environmental, social and economic challenges we face. But they can contribute best by continuing to find ways for deep-dive investigations on one hand and integrated, multi- and trans-disciplinary research on the other. Ultimately public science is paid for by citizens and we need to ensure that there is a sense of civic ownership and engagement with what we do. But on the other hand much of science, and especially technological development, also occurs in the private sector and we need to ensure that barriers are not inappropriately erected while recognising the different interests of different stakeholders and managing any conflicts.
The more utilitarian framing of the role of public science in recent decades reflects the close relationship between science and the rest of society. This framing has emphasised that science has essential value for improving the human and planetary condition and for driving innovation and economic growth. And indeed it is these arguments that have driven many countries to increase their investment in public science and to promote private sector R&D. But the importance of discovery science cannot be underestimated. Indeed the history of science is full of spinoffs from basic enquiry. At the same time, as Daniel Sarewitz has pointed out, science has been accelerated in many of its discovery efforts by technological advances that have made new ways of investigating possible. Ensuring the balance of different modalities of science is essential.
Over the past decade or so I have come to appreciate the view from ‘in-between’. Assisted by scholarship in the humanities and social sciences, which have been observing and studying human knowledge production for decades, and by the advent of complexity and systems-based approaches in research, I have learned to question tidy answers to complex problems. It is in the interstices that we can find out what is really going on, which may offer our best hope for successfully addressing the complexity of our health, social and environmental challenges.
Whether it is between disciplines, sectors or jurisdictions, I believe what is needed is greater attention to brokerage. Science and its context (policy and society) shape one another in continuous and mutually refractive processes. Those of us at the interface of these processes are positioned to broker the insights between science and policy communities, to help engage with the broader society, and to understand how each of these domains influences the others.
Importantly, we also have a responsibility to understand and make known the limits of science-based knowledge. As Paul Cairney in his book The Politics of Evidence Based Policy Making points out, at the very time that policy makers need inputs, the science is generally incomplete and often ambiguous. He goes on to point out the reality that while (disciplinary) science is often very good at problem identification, it often is not well placed to find policy-acceptable solutions – here I think greater interdisciplinarity and engagement with policy communities is important, hence the importance of a brokerage approach.
The broker’s role is to help identify what we know from science, what are the limits and caveats on that knowledge, what remains unknown and what are the implications of that knowledge for the options that the policy community must choose from. In the science advice ecosystem, we need knowledge generators like universities and research institutes (and funding institutions to support research), knowledge synthesisers like academies, associations and unions, institutions that protect science like unions and academies and formal knowledge brokerage structures like science commissions, advisory committees and science advisors.
But the closer the broker is to the policy community, for them to be effective it becomes more important not to pronounce on values-based judgements which are rightly the societal discussions to be led by policy makers and politicians in a democracy. Instead, science advice should be present and fully engaged to help support their discussions. The distinction is subtle but significant. Other actors within the science advice ecosystem such as academies must have more flexibility in this regard, which is why the various roles in the overall ecosystem are complementary.
Academics as individuals must be fully protected in their roles and their expertise, as academic freedom remains a critical principle that is to the ultimate benefit of society.
I want to thank the many people who have helped me over the past nine years, and in particular the past and present departmental science advisors and my own team. I would particularly want to acknowledge the contributions of Prof Stephen Goldson, Dr Alan Beedle, Dr Anne Bardsley, Dr Felicia Low, Dr Tatjana Buklijas, Kristiann Allen and Megan Stünzner in OPMCSA and Lara Cowen and Grant Mills in the INGSA office. There are many other New Zealand and overseas colleagues who have been an enormous support in sharing insights and testing ideas, for whose collegiality I am grateful.
I now will be putting more effort into the further development of the International Network for Government Science Advice (INGSA) which has grown rapidly over the past four years and now has over 5000 members in about 80 countries with chapters in Africa, Latin America, Asia and developing chapters in Europe and North America.
I will be helping to establish a centre, within the University of Auckland’s Public Policy Institute, which will focus on science-policy interface. We are going to call it SciPoDS – the Centre for Science in Policy, Diplomacy and Society. I will also be continuing to assist our Ministry of Foreign Affairs and Trade and, in particular, continuing to act as the coordinator of the Small Advanced Economies Initiative.