A government data policy deadline is nearing. Data leaders are looking to help

At the end of September, federal agencies will be required to publish a comprehensive inventory of their data assets on their websites, ensuring that those entries meet a new metadata schema as well as open formatting requirements.

The Sept. 30 deadline will be a milestone for the data quality and accessibility improvements envisioned in a 2019 law known as the OPEN Government Data Act, capping a list of actions the Office of Management and Budget outlined for that implementation in the final days of the Biden administration. 

In an effort to help make part of that transition easier, the Chief Data Officers Council is sharing their homework.

At the end of August, the CDO Council published a 78-page implementation guide that aims to assist agencies in their efforts to comply with new standards for federal metadata that were published earlier this year. Put simply, those standards — the DCAT-US v3.0 schema — updated how the federal government goes about describing its data to make it more accessible for modern tools. 

“The government cannot exist without data, but you can’t protect it if you don’t know that you have it. You can’t use it if you don’t know that you have it. AI readiness is nearly impossible if you don’t know what you have and what it can or cannot be used for,” CDO Council Chair Kirsten Dalboe told FedScoop. “So an initiative like this helps a lot of things.”

The guide is focused specifically on the requirements under the OMB memo related to metadata for the inventories and submission to Data.gov. It includes a five-step implementation approach and spans topics such as data asset identification, best practices for non-public data assets, and updating data inventories. It also includes specific guidance for geospatial data and examples of data structures and governance. 

While agencies were previously required to publish an inventory of publicly available data assets under early open data requirements, the 2019 law transformed those public postings into comprehensive inventories that include documentation of items that can’t be made public.

“It’s frankly a change in process for many agencies where they maybe have not been doing as much cataloging on the non-public data, and so now we have to rethink that and how we go about doing that,” Dalboe said.

As a result, the guidance goes over issues such as how those disclosures intersect with cybersecurity practices and how to prioritize when filling the inventories out.

On the security front, Dalboe said the new inventory requirement creates “an interesting nexus” as it enters into the area of zero trust — a cybersecurity principle in which every user or device is treated with skepticism and verification is constant. That intersection raised questions including what the sensitivity level was for data assets, Dalboe explained.

“If we’re saying something’s non-public, we have to give a reason why. So that was part of the complication of, ‘well, how do I go about doing that?’” she said. 

Meanwhile, a phased approach strategy in the guidance is an effort to acknowledge that achieving a comprehensive inventory will likely extend beyond the Sep. 30 deadline for many agencies.

“The idea that every single data asset will be magically inventoried on September 30th will probably be very difficult, so let’s think about a phased approach so that every agency can show progress towards this requirement — recognizing that, probably, we will never all be done,” Dalboe said. “There’s always more data to catalog.”

Data in AI era

The guidance has earned praise from data policy advocates.

Taka Ariga, vice president of data policy and impact at the Data Foundation, applauded the inclusion of JSON examples — a text format that’s used to structure data behind the scenes — that shows agencies what they might want to implement.

Ariga said that’s important for consistency but more importantly because it’s directed at the implementers rather than coming from a strictly policy view. Often, he said there can be a disconnect between policy and actual implementation. Ariga was previously a federal data leader at the Government Accountability Office and the Office of Personnel Management, and was a member of the CDO Council. 

He also highlighted the importance of the schema changes in the era of AI. While it’s great that the data is being cataloged and inventoried, “in the AI world, I think context reigns supreme,” Ariga said.  

For example, in government “CDC” could mean the Centers for Disease Control and Prevention, or calculated date of confinement. The policy changes help tools like chatbots have the right context to avoid AI slop, Ariga said. 

Ariga said the guidance is the CDO Council’s way of addressing how to lower the barrier toward implementation by providing a blueprint that’s easy to follow. 

If successfully implemented, Ariga said “I think this will drastically change in an almost invisible way how much greater accessibility that the open data will” provide across AI platforms, whether that’s Gemini, ChatGPT or Claude.

Slimmer workforce, same requirements

Notably, the guidance and upcoming deadline come after major upheaval among the federal workforce under the Trump administration — including among data teams. 

Specifically, the Data Foundation’s 2025 survey of agency chief data officers found that two-thirds reported changes in their staffing support over the past year, with nearly 40% reporting losses of six or more staff members. Some CDOs even expressed concern about their ability to implement the requirements, such as the OMB memo, per the report.

The loss of staff is compounded with a delay in the implementation timeline. While the updated schema was initially expected to be released in July 2025 under the memo, it was published in May of this year, leaving agencies with less time to prepare for the September deadline than originally envisioned.

Lighter staffing loads as well as the loss of certain technologies were among the factors the council considered in creating the guidance, Dalboe said. 

Some agencies, for example, lost the data cataloging tools that they would have used to meet some of the requirements, she said. “There were also discussions about ways to comply if you no longer have a tool to do this and need to shift into Excel, for example.”

In addition to the CDO Council, representatives from Data.gov and the General Services Administration also contributed to the project. According to Dalboe, it’s been in the works for several months.

The CDO Council started by asking agencies to lend technical support to the project — for example, the ability to read JSON code and use GitHub — and later broadened input to a more general group. Each week, Dalboe said about 10 to 15 people would get on calls and work through how to resolve a list of questions — from bigger picture to in-the-weeds technical issues.

“It was a pretty big effort, and just really exciting to see so many agencies and communities come together to figure this out,” she said.