For scientists, thoughts about democracy can feel distant from their work. Yet, around the world, governments are limiting academic freedom and interfering with decisions about what research is funded and who performs it¹. Some scientists are also facing politically motivated interference and attacks².

Six principles for supporting academic freedom in universities

As members of the European Commission’s Group of Chief Scientific Advisors, we feel it is urgent to address this concern and that we have a responsibility to make scientists’ voices heard.

Science and democracy are often viewed as distinct — science as a source of discovery, expertise and innovation and democracy as a set of institutions, rules and procedures. But, in reality, the two are intertwined³.

Science is an essential backbone of democracy. In an increasingly complex and fast-changing world, people need access to verified information and independent knowledge to make informed decisions^4,5. Research aids that by generating reliable evidence and fostering critical enquiry, thus enabling a plurality of views to be discussed and creating conditions in which societies can learn from experience.

In May, Ekaterina Zaharieva, the European Commissioner for Startups, Research and Innovation, asked us to provide advice on how science can strengthen the European Union’s democratic functioning and resilience.

Independently of that process, here we put forward three pillars through which science enables democracy: establishing shared factual baselines, aiding trust in institutions and providing tools to adapt policies and organizations in response to changing circumstances.

Shared factual baselines underpin democratic decision-making

The first pillar relates to protecting the foundations and integrity of information on which democratic judgements, collective deliberation and, ultimately, action depend.

Democratic societies constantly confront questions for which there are no simple, best or permanent answers. Scientific institutions help societies to navigate changing realities by creating procedures for gathering evidence about the current situation, past trends, probable outcomes or management options; assessing impacts; communicating disagreement; quantifying uncertainties; and updating conclusions in light of new findings.

Artificial-intelligence and other digital platforms are changing how people find information, develop opinions and take part in public life⁶. This not only provides opportunities for broadening participation and increasing transparency, but also produces intensified, increasingly complex or fresh risks, including of misinformation, manipulation, bias, polarization and fragmentation^6,7. Moreover, these developments amplify the disproportionate influence of a small number of powerful technology companies and their leaders⁸, including their ability to ‘manufacture doubt’ and shape public opinion^9,10.

How to get science back into policymaking

The challenge goes further than just correcting people’s misconceptions. Disinformation is often rooted in social networks, reinforced by emotions and group identities and used deliberately for political purposes^5,7. It spreads through networks of people connected by identity, grievance or distrust. These vulnerabilities can be exploited by individuals seeking to advance political or economic interests⁹, thereby undermining democratic institutions, deepening divisions and fostering instability⁶.

Citizens have legitimate concerns about economic restructuring, regional inequities, how migration is governed, technological changes, ageing populations, climate risks and geopolitical instability. People might lose trust in democratic institutions, especially if they feel they are not being heard, do not understand why something changed, think they are treated unfairly or see the fabric of their communities being disrupted¹¹. This might create vulnerabilities that can be exploited through disinformation and political manipulation⁵.

For this reason, scientific approaches to democratic resilience must go beyond fact-checking and include the design of effective channels for science communication and information sharing, as well as opportunities for meaningful public participation and informed deliberation^4,12. These should include transparent data sources in digital platforms; independent media; strong safeguards against foreign interference; secure information infrastructures; independent monitoring and assurance of algorithmic systems; and improved understanding of how the expanding use of AI tools might shape democratic discourse and outcomes¹³.

Science-informed governance can strengthen trust in democracy

The second pillar refers to the legitimacy of expertise and the trustworthiness of institutions. Democratic governance can be thought of as a shared societal capacity to understand complex realities, to deliberate across differences and to act together in the face of social, economic and technological transformations and uncertainty^4,6. Its resilience rests on organizations that sustain public trust and active citizenship, and on people’s capacity to participate meaningfully in deliberation and collective problem-solving^4,14.

Institutions that are informed by scientific evidence and guided by scientific principles are better equipped than others to make effective, transparent and accountable decisions, strengthening public trust in democracy¹¹. During a pandemic or other crisis, or in anticipation of such events, authorities can reinforce trust by explaining clearly how scientific evidence, including any limitations and uncertainties, informs their decisions.

But for science to strengthen democracy, researchers must be trusted, too^15,16. Moreover, the legitimacy of expertise has become increasingly contested over the past few years in Europe and elsewhere^5,15.

Scientific advice is subjected to politicization in many countries, with consensual insights and clear evidence (such as on climate change) being dismissed as ‘just another opinion’ or as being driven by a political agenda^5,9. Public trust in expertise has become polarized along political lines^5,15. Some officials and politicians might support specific strands of science, for instance, by promoting individual researchers or a few disciplines selectively. Such favouritism has implications for funding.

Six ways to put the public at the heart of science and policy

This ‘divide and rule’ approach to science is often tacitly accepted by the academic community, with many institutions preferring to stay silent rather than run the risk of their budgets being cut. The dilution of scientific knowledge in an ever-increasing flow of online information, often promoted by individuals with vested interests seeking to influence public decisions, can also render impartial scientific evidence and advice less visible.

Expertise strengthens democratic legitimacy when it is transparent, diverse, independent and open to scrutiny — as prescribed by the rules of science. But it weakens legitimacy when it is perceived as opaque, unclear, politicized, elitist or disconnected from people’s lived experiences.

Ensuring that expert information is democratically legitimate requires institutions and mechanisms that not only structure the interaction between science, policy and society, but also protect organizations’ separate, independent roles, manage conflicts of interests and prevent undue influence.

Crucially, not every engagement with science carries equal weight. Impartial access to scientific knowledge is most reliably secured through a cross-disciplinary and diverse group of experts. By contrast, the opinion of a single scientist might result in partial understanding (highlighting a particular disciplinary perspective) or reliance on a dissenting voice that differs from the majority view. High-level expert groups, which engage directly with senior decision makers, can identify a broader range of potential concerns. They can better understand the wider implications of policy challenges to provide relevant scientific evidence and advice, yet maintain independence in their analysis, conclusions and recommendations.

Have people stopped trusting science? The data tell a surprising story

Public institutions should ensure that scientists who provide advice can do so independently and that they are free to publish their findings. This generally requires stable public funding of research¹. Cuts to public science budgets can increase scientists’ dependence on a narrow range of funders, including private and commercial organizations or institutions linked to other countries. Their funding decisions could reflect strategic or political interests and might create actual or perceived conflicts of interest if appropriate transparency guidelines and safeguards are not in place.

The European Commission’s science-advice framework provides a useful example of how scientific independence can be achieved. It combines topic-specific evidence reviews by multidisciplinary expert groups with independent policy recommendations from its seven chief scientific advisers. The advisers are supported by a permanent secretariat that helps them to align scientific advice with policy needs and safeguards the independence of the advisory process. This mechanism also includes rules on transparency and the disclosure of advisers’ conflicts of interest (see go.nature.com/4jndrg).

Scientific expert groups should be able to propose topics, identify questions and communicate evidence, uncertainties and policy options to decision makers and the public independently.

But trust depends on people having the opportunity to engage with science, too. Scientific literacy, civic education, public engagement with research and access to reliable information are all essential and need to be supported by public and educational institutions. Opportunities for science education are a public good and should be accessible to all.