The more than 60,000 scientists who work in the U.S. government form one of our nation’s greatest assets. Their research is as varied as science itself, ranging from understanding the fundamental workings of atoms to building the next great space telescope to determining unsafe levels of chemicals in waterways to predicting the impacts of future climate change to creating new vaccines and cancer treatments.

We have a clear interest in knowing what the scientists we support are doing, both as a matter of public accountability, and so that we can make informed decisions. While we can obtain this information in a variety of ways, including journal articles and information on agency websites and social media accounts, I would argue that no channel of information about government science is more critical than the independent media.

Consider, for a moment, the word “media.” The media mediate between sources of information and the public. They make decisions about what is most important and relevant to readers (since no one has time to keep up with all the science being done at federal agencies, or even one agency). They translate from the technical to the accessible. They place science in larger social and political contexts, and they hold institutions accountable when they try to manipulate or suppress scientific results for political reasons. No other institution in our society is capable of fulfilling all these roles. For this reason, free and open access to government scientists must remain open, even—indeed, especially—when scientists’ results challenge the government’s political outlook.

How open are federal agencies to the press now? The answer—at the moment, at least—depends greatly on the agency. At some, such as NASA, NIST and NOAA, a reporter can in most cases speak to a scientist without having the interview cleared in advance with a public affairs officer, or PIO. PIOs at these agencies generally help reporters connect with scientists, and provide backup information to support a story.

At other agencies, however, staff put up substantial barriers between journalists and scientists. The EPA, for example, requires many interview requests to be approved by PIOs. This may make sense in some cases, as the EPA is under constant pressure from both anti-regulatory and environmental groups, and wants to tailor its message to avoid costly political blowback. But that does not get it off the hook for providing timely scientific information to the public—and delays in approving interviews can kill stories being written on tight deadlines.

Access barriers can also pop up for seemingly innocuous topics. When I contacted a U.S. Forest Service scientist a few years ago to learn about research being done to understand threats to the eastern hemlock tree—hardly an area of major public controversy—I was shocked to get a reply from a PIO asking for a list of questions I planned to ask. (I did eventually talk to the scientist, but only after several days’ delay.) I have had a few similar experiences at other agencies, and have heard and read far more egregious stories from others.

I will grant that on matters of policy, an agency has an interest in speaking with one voice, so as not to confuse the public. But when it comes to the science informing policy, the public deserves the unfiltered and unmanipulated truth, directly from the scientists who did the research. It’s not just I who believe this. The Union of Concerned Scientists has for years advocated for openness at government agencies, and its 2015 “report card” on agencies’ media and social media policies reported an overall improvement over the Obama years. “Scientists now have the expectation that they’ll be able to share their research and their views with the public,” says Michael Halpern, deputy director of the union’s Center for Science and Democracy, “and our democracy is better for it.”

But the report found many agencies’ policies still falling short of ideal. (One of the agencies to which the union gave an “incomplete,” the Department of Energy, implemented a strong openness policy in the waning days of the Obama administration. Let’s hope the new administration honors it.) Meanwhile the Society for Professional Journalists wrote several letters imploring the Obama administration to live up to its promise to be the most transparent administration ever. (I chair the National Association of Science Writers’ information access committee, which has supported several of these letters.)

I can almost hear you getting impatient at this point. “What,” you’re asking, “is information access at federal agencies going to be like under Trump?” As with so many things, the incoming administration’s plans in this area are still largely unknown. But early signs are troubling. Confusingly worded directives led to temporary freezes on communications at several agencies, though some of these have been reversed after a public outcry. (On the plus side, these events showed that the public is engaged on this issue, and that pressure can have an impact.) At the moment, sources at various agencies tell me it’s mostly business as usual, but those agencies are still awaiting confirmed nominees. And it’s abundantly clear that Trump himself is no fan of the press, to put it mildly.

Those who wish to explore the darkest possible future might read the American Heritage Foundation’s 2016 “Blueprint for Balance” document—believed by some to be the basis of President Trump’s budget plan—which suggests saving $262 million (.007 percent of the total federal budget) by eliminating public affairs staff at agencies. This would, of course, not stop journalists from finding their ways into agencies, but would probably make their work more reliant on scientists willing to leak or go outside authorized communication channels, and would certainly make the American public more ignorant about the workings of its government—and about science.

Even without this “nuclear” option, there are plenty of ways that administrators can restrict access to scientists and information. They can ignore or delay responses to inquiries in hopes of delaying or killing a story. They can route interview requests to scientists they know will toe the party line. They can try to suppress individual scientists’ ability to use social media, speak at conferences, and, in the worst case, publish research without review. Unfortunately, Donald Trump’s team would not need to invent these tactics, because they have already been demonstrated by his predecessors. They would merely need to make them the rules instead of the exceptions.

As the new administration’s media access policies become clear, journalists and the public must be vigilant to ensure that scientific integrity and free flow of information remain enshrined as policy and practice across the federal government. These principles are vital to our democracy.