The university I lead is more than four centuries old and has seen political orders rise and fall. I have not, but I work each day in an institution that remembers those changes — in its archives, statutes and culture of argument. The conditions that make knowledge trustworthy, such as independent peer review, openness about methods and data and the ability to correct earlier work, are changing in ways that reach the core of what a university is for.

How universities came to be — and why they are in trouble now

Those conditions are under pressure in a growing number of countries, and the familiar tools of quiet diplomacy, internal negotiation and isolated warnings no longer keep up with the pace of change. For example, in the past decade, governments in various countries have narrowed the space for independent evidence in different ways.

In the United States, changes to research-funding mechanisms and programme priorities in areas such as climate science and gender research, and the sidelining of environmental and public-health expertise, have shown how quickly political decisions can affect the systems that produce independent evidence (see go.nature.com/4hgg3ym and go.nature.com/4zzmtur). European academies have voiced their concerns in public statements on the pressures facing US science and international collaboration^1–2(see also go.nature.com/4a3pn9s).

In Hungary, a sustained legal campaign around foreign higher-education institutions drove Budapest’s Central European University out of the country to Vienna in 2019. In 2020, the Court of Justice of the European Union found those legal changes incompatible with EU law^4,5. The political landscape in Hungary has shifted since then, but the legal tools used and the damage to institutional autonomy and academic freedom remain instructive.

Elsewhere, grant language is being rewritten to steer funding priorities towards favoured themes such as economic competitiveness, innovation and national security. Access to public data in sensitive areas such as climate and health is being reduced, and researchers are adapting proposals accordingly, sometimes by avoiding or adding certain topics or terms to maximize their chances in the next funding round.

The details differ, yet the pattern is familiar to many of us who work in universities. One question keeps returning: when governments or funders intervene in universities, do they strengthen or weaken the checks that make knowledge trustworthy?

Others have warned about these threats to independent enquiry already^1,2,6. Here, I try to turn these concerns into a set of principles that can guide real decisions in the messy situations academics face. They can be used as a test for any measure that affects universities, and apply to governments, funders and university leaders alike.

They are written for people who have to decide, often under time pressure, where legitimate steering of a publicly funded system ends and interference with independent knowledge begins. They concern both the freedom of academics to pursue questions wherever the evidence leads, and the ability of universities to protect that work from undue influence. These are distinct but connected responsibilities, because pressures on one often weaken the other.

The first three principles set out why independent enquiry must be protected and how that protection is anchored in the university’s core task. The last three show when that protection has to be weighed against other legitimate concerns.

Ground academic freedom in the duty to produce reliable knowledge

Universities have a responsibility to produce reliable knowledge of lasting public value — including for people who will never set foot in a lecture hall — and to teach students how to think using that knowledge. Academic freedom is the working condition that makes this possible.

I learnt this from supervisors who allowed me to ask awkward questions and from students who pushed back in seminars. We are answerable first for the quality and integrity of the knowledge we produce and teach. Academic freedom is one of the tools that allow us to live up to that responsibility.

Discussions about academic freedom often start with rights and only later bring in responsibility as a limit. I prefer to reverse that order: universities are first accountable for the quality of the knowledge they create and pass on, and academic freedom and institutional autonomy are the conditions in which to do this well.

Research that bends towards political or commercial pressure carries that pressure into the questions that are asked, the methods that are chosen and the outcomes that are presented^7,8. Teaching that pushes political instructions to the fore trains students to adapt their thinking to approved answers instead of testing arguments against disciplinary standards (see go.nature.com/4a3pn9s).

External directives that decide in advance which questions might be asked or which answers are acceptable reduce the space to question assumptions and to correct mistakes. When the people who produce and teach knowledge cannot work independently, they gradually lose the ability to correct each other and to correct those in power.

Quality is often the main reason given for defending academic freedom. It is also the main language used to justify limiting it. Some governments frame measures that reduce universities’ capacity for independent enquiry in terms of quality, integrity, balance or public safety, presenting these interventions as ways to protect academic standards. What matters is how knowledge is produced in practice, whether others can see and test what has been done and whether it proves reliable over decades.

Measures that genuinely improve knowledge make methods and results easier to check and to challenge. However, changing grant calls to favour certain answers, screening researchers for their views or taking uncomfortable data offline weakens that scrutiny and minimizes what can be known from work that has already been done. The test for any such measure is whether it strengthens independent scrutiny, and with it the university’s ability to meet its responsibility to society.

Protect questions that have value that cannot be shown in advance

Curiosity-driven research often leads to applications that no one can predict. If researchers limit in advance which questions they will pursue, they also limit the futures that science can open up.

There are various examples of such applications that were initially unpredictable. Work on how bacteria defend themselves against viruses became CRISPR gene editing, now the basis of a therapy for sickle-cell disease. Research in linguistics and digital humanities helped to build translation tools and search systems. Long-term epidemiological cohort studies underpin public-health guidance on smoking, diet, workplace safety and environmental exposure. Decades of work on messenger RNA and lipid-nanoparticle delivery became the platform for COVID-19 vaccines when the world needed them. None of these lines of work would have scored highly for societal relevance at the start.

Key US science panels are being axed — and others are becoming less open

Research should speak to societal challenges, but funding systems also need to preserve room for questions with a value that cannot yet be specified. Narrowly defined relevance criteria tend to reward safe, incremental projects and push researchers towards low-risk enquiry. Free enquiry matters precisely because scientists do not know in advance which questions will turn out to be important.

A government that wants to base policy on robust knowledge depends on a research system that can generate and test that knowledge independently. Cutting public investment in fundamental research might reduce spending today, but it shifts the cost into the future — through missed treatments, weaker prevention, lost innovation and a reduced capacity to respond when the next crisis arrives. The test is whether a measure unduly narrows the range of questions that can still be asked.

Treat universities as democratic infrastructure

Universities disperse power in practice by training professionals who can question official decisions, by producing analyses that can contradict government or corporate claims and by placing independent expertise inside courts of law, in the media and in public debate. Together with the executive, legislative and judicial branches, and with a free press and civil society, universities form a pillar of a society that is built on democracy, human rights and the rule of law. When that wider structure is weakened, universities are usually affected as well.

Hungary shows how this can unfold. Between 2017 and 2019, Viktor Orbán’s government reduced the independence of universities and research institutes by transferring control of research funding from the national academy of sciences to a new research network. The network’s board is chaired by a government-appointed minister’s nominee. The government used higher-education law to force the Central European University abroad, and brought law courts, the press and civil society under tighter political control through parallel legislative changes⁴.

Elsewhere, earmarked funding, new laws on cooperation, politicized appointments to governing boards and a steady public attack on expertise point in a similar direction. Examples of such trends are described in global and EU-wide academic freedom monitors and national reports^9–11.

It does not take dramatic events abroad to damage academic freedom. In my own country, the Netherlands, the national academy warned in 2025 that large budget cuts increase the pressure to shift from independent to commissioned research, and that a proposed screening law on knowledge security would limit researchers’ freedom to choose their collaboration partners. It also documented self-censorship after researchers were intimidated in public debates¹². Rectors of Dutch universities warned in the same year of a growing political tendency to steer the content of scientific debate and teaching (see go.nature.com/3v3ktjj).

Institutional autonomy means being accountable to the public for the quality and integrity of knowledge, rather than to a government for whether it finds the conclusions politically acceptable. Rules and accountability frameworks are not in themselves a problem. Many exist to protect people, ensure workplace safety or guard public funds. A problem arises when their combined effect, including funding conditions, reporting requirements and rules on cooperation, shifts the centre of control from the university to the ministry. One form of accountability strengthens independent knowledge; the other replaces it with political acceptability.

Hold security and openness as a single responsibility

The case for academic freedom is linked to that for research security. The risks of espionage, loss of intellectual property and harmful use of sensitive knowledge are real, and universities cannot ignore them.

At the same time, security measures that are designed without attention to academic freedom can unintentionally narrow independent enquiry and place a great deal of control in the hands of those who manage access to data, infrastructure and international collaboration.

Scientists fight back against far-right plans to restrict academic freedom in Germany