Blog

Nick Hart There’s a lot of great interest in data across the country right now, and for organizations like us at the Data Foundation where we’ve been championing open data and evidence-informed decision-making for a decade, this is an interesting moment because a lot the important discourse that’s happening around data is actually very similar to the discourse we’ve been having for years. One of the key themes that we’re watching is how data that government has already collected are actually being protected. Are we abiding by the laws, the privacy restrictions, the project approvals that are necessary? I think that’s a lot of what we’ve heard from some of the folks that are particularly concerned about the Department of Government Efficiencies actions, which is to say, they may not be following all of those protocols. However we look at this, though, it’s really important that any administration always protect the American people and the American businesses’ privacy of any data that’s been collected. And we need mechanisms to also understand how the data are being used. So that’s sort of the second prong here, which is government has never been particularly good at sharing data use. And this administration, the Trump administration, has pledged radical transparency to communicate data uses. I think that’s something that we wholeheartedly support, but we’d also like to see better mechanisms to do that.

Terry Gerton Well, one of the things that you and I and many others in the good government space have advocated for is more cross-agency, cross-silo connectivity of these databases to get a better handle on customer service or who’s in the system and how they’re receiving it. The DOGE team seems to have broken down a lot of those barriers. Some of us would argue those barriers were there for a reason, but now they’re down. So what happens now? Are we going to get to a point where we can really share data better?

Nick Hart Well, hopefully we can get to a point where we can share data better, but also responsibly and encourage ethical uses. And this was the underpinning of a strategy that came out of the first Trump administration called the Federal Data Strategy. Actually, Suzette Kent, the former federal CIO, and I published an op-ed here with federal news network, back in December of 2024, that called for a refresh of that strategy, which never imagined in the current world that we’re operating in, including things like AI that have become very contemporaneous in our activities in government. So currently, the administration has not leaned in very far on refreshing the Federal Data Strategy, but we just met with the Office of Management and Budget recently and hope to be able to see that refreshed soon. The second thread here is really just to say that the executive order the president issued on breaking down barriers to data silos, I think that’s actually a really important thing for the country to discuss. Much of the detail in that executive order was actually similar to things that we’ve been saying for decades. We need to enable data sharing. We needed to encourage the use of the data the government’s already collecting for a variety of purposes. Now this administration’s really leaned in on reducing fraud or eliminating fraud and improper payments. And that’s one use of data. But there are a lot of other uses of data that are really important in our society. Knowledge generation for research. The business community has a great need to use government data for investment decisions. We all use data from places like NOAA, the National Oceanic and Atmospheric Administration, to understand whether it’s going to be hot today or a thunderstorm later so you can bring your umbrella. So however you look at those uses, you need the security of the information, the privacy protections must be strong, and ultimately the information has to be accessible.

Terry Gerton So what would a refresh of the data strategy look like and sound like in practical steps? What would you be looking for there?

Nick Hart The data strategy originally included 10 principles and 20 broad themes for implementation, and then it was envisioned to have these annual action plans basically continued all the way through the Biden administration. I think the first step to realize is that we actually do need annual action plans to come through, to give the chief data officers that now exist in every federal agency, as well as other data leaders around government direction, about what to do. A refresh would also really encapsulate some of the core components of a law called the Foundations for Evidence-Based Policymaking Act that you and I have talked about many times before. It included a subset of directives around open data and protecting of statistical information in the federal statistical agencies. A refresh will really double down on the need to explain to the American public what data exists that government currently has. Our inventory and cataloging infrastructure is incredibly weak. So if you’re a person on Main Street out there, my parents back in Missouri, you really have a hard time figuring out what data government actually collects and possesses about you. So for those in the current environment that are interested in minimizing information about themselves and government databases, it’d be really hard to figure out how to do that. So, enabling a mechanism to better track what data actually exists. But also things like when we’re terminating a data collection — which has been a major topic of some of the DOGE themes recently — which is to say, if we’re no longer going to collect a particular data set at the Bureau of Labor Statistics or the Bureau Economic Analysis or any other agency, how do we communicate that to the American public to ensure there’s good engagement between the users of that information and government itself?

Terry Gerton I’m speaking with Nick Hart. He’s the president and CEO of the Data Foundation. Nick, I want to go back to a point that you mentioned before, which is statistical data. You’ve recently expressed some concern about the reliability of federal statistical data, let me ask you to do two things. First, talk really quickly about the federal statistical agencies and what kind of data they collect, and then what are your big concerns?

Nick Hart Yeah, so the federal statistical agencies are really the heart of the country’s data collection and management infrastructure. They collect some of the core information that we need to understand the population and the economy. The Census Bureau, well-known, every 10 years they conduct a census and that has information on every person in the United States. The Bureau of Economic Analysis produces something called the gross domestic product. The Bureau of Labor Statistics has the unemployment rate. Across the 13 major statistical agencies dispersed across government, they collect and manage a whole host of information about the American people and American businesses. Back in May, the Data Foundation learned about a survey that DOGE was conducting of the federal statistical agencies, in particular data collections. And we thought this was a bit peculiar, in part because decisions typically about statistical activities are protected due to requirements for independence from the political processes. And that’s not to say that we don’t use values and different perspectives to determine what information to collect and how to manage it. But typically, the leaders of statistical agencies or the chief statistician of the United States would be making those determinations about both what to collect and what not to collect, and when to end terminations. We know currently in the environment today, and the Data Foundation just published something on June 30th, that indicated there are a number of federal statistical agencies that have been reducing data collections because of attrition in staff and also limited resources. So that includes things like reducing the ways that we collect the CPI, the consumer product information index, for inflation. And that will have potentially some impacts on users around the quality of the information in the years ahead.

Terry Gerton And that kind of data is essential. It’s essential that it be trusted and transparent in order to allow the markets to function effectively, right?

Nick Hart That’s exactly right. Actually, one of the pieces of the Foundations for Evidence-Based Policymaking Act, or Evidence Act, required a regulation to encourage public trust in statistical data. That regulation was put out for public comment during the Biden administration, it became final, but today it’s foundational and helping to protect statistical data. It is in effect today. And we’ve seen no measures, by the way, that the current administration is going to walk back from that. But for the entire community of data users out there, it’s something that we can lean on to say, these data are so important to the American people and businesses that we actually need to make sure that they’re protected as well as have fidelity to the independence of the system.

Terry Gerton And so as you think about this trusted and transparent data for market, effective market information and decisions, it seems to me that this administration is also kind of presenting AI as its way to fill in the gaps for the people it no longer has, but AI has to be trained on trusted, transparent data as well. So what are you seeing in that space?

Nick Hart A lot of AI systems and large language models have incorporated open data from the federal statistical system, as well as other parts of government to train. And that’s one of the great benefits that government data can have right now for AI. But we also know we need quality information to be part of those training sets. So just because you have data doesn’t mean it’s good data. And this has been one of the longstanding challenges as we’ve been talking about AI — is that it requires an immense amount of knowledge to train AI models. And wherever that knowledge is coming from, whether it’s the private sector or the public sector, you need to be able to have some embodiment of trust in that information. And again, this is why the federal statistical system is so core in moving forward. They’ve really demonstrated approaches for transparent data collection and management, as well as enabling high-quality information to be available for the American people.

Terry Gerton So you’re sitting at the center of this conversation around what data we collect, what data were going to share, how we’re going to collect that data. What do you see happening in the near future? How is this conversation playing out with the Trump administration?

Nick Hart Well, there’s probably more public discourse about data right now than at any point in my life. And I find that really exciting, though I realize that for some that is probably a bit of trepidation that they’re living and experiencing. But I would say this administration has really leaned in on using data to make decisions. And that may be a controversial statement to some listeners. However, whether it’s the DOGE activities or other things that we’re seeing happen across the federal environment right now, there’s a real discussion about how we better use data for decision-making. I think one of the great challenges that we’re going to face that we maybe haven’t seen as much in recent years is more discourse about the uses of the data, including what areas we should be encouraging data minimization for. So that’s a concept that means we may not actually need to collect as much information as we once did. It’s one of hopefully great benefits of having improved AI in coming years: that we can better understand what data is the most meaningful for measuring program impacts and outcomes or understanding the population and the economy, as well as improving areas like basic data standards that help us harmonize and create interoperable systems. So instead of asking your income in 12 places, Terry, we could ask it once and have a protected data set that helps reduce burden on the American people.