AI is a rising priority for federal chief data officers
Nearly all CDOs surveyed by the Data Foundation reported thinking about adopting the emerging technology.
In 2019, the passage of the OPEN Government Data Act as part of a larger evidence-based policymaking bill required many of the largest federal agencies to establish chief data officer positions, responsible for managing their agencies’ data assets and carrying out certain statutory requirements.
Years later, the Data Foundation’s annual report, done in collaboration with Deloitte, surveyed 27 department-, agency- and bureau-level CDOs and statistical officials on their primary focuses and challenges.
One big priority for CDOs this year: artificial intelligence.
Just above half of CDOs report already using AI, and 95% reported that they are thinking about adopting AI in their organizations in the next year.
That includes Rob King, CDO at the Department of Energy and incoming vice chair of the Federal CDO Council.
King said at a Tuesday event held to release the Data Foundation report that he is looking into using AI and automation technologies to take on some of the data stewardship activities — like data classification and tagging — that have traditionally been underinvested in. The work is foundational to preparing data for use with AI.
“Advanced analytics, master data management, data integration, API strategies — all are going to help position those building blocks for really leveraging and driving ethical and explainable AI,” King said.
The Data Foundation argues in the report that more clarity about how the CDO role relates to AI is needed.
Draft guidance on AI issued by the Office of Management and Budget last month included a provision to create new chief AI officers in agencies, charged with coordination, innovation and risk management related to the emerging technology. Those new officials could be dual-hatted CDOs, chief information officers or other relevant officials, the guidance said.
Another CDO-focused survey released earlier this month — done by the Federal CDO Council — found that 40% of its 35 agency CDO respondents were also the primary accountable official for AI, with another 45% saying they were a partner in the area. That’s up from 25% on both answers the year prior.
Another recommendation from the foundation is the development of “clear, ethical guidelines and governance frameworks to support CDOs in responsibly adopting emerging technologies like AI in service of their public mission.”
Where CDOs sit in their agencies’ org charts still varies, according to the latest survey, with about a third reporting to CIOs and others reporting to positions like agency heads or chief operating officers.
The Data Foundation also recommends clarifying authorities and responsibilities for CDOs, especially in relation to agency CIOs. The report suggests OMB issue guidance on the implementation of the OPEN Government Data Act, with specifics on roles and responsibilities of the CDO — an echo of its call for OMB to release that guidance last year as well. The law is intended to improve the availability and transparency of government information.
Other recommendations from the Data Foundation include more resources, staff and training for CDOs.
“We found resources, skills and authority are persistent barriers. This is consistent with all of our past surveys,” said Katie O’Toole, senior policy and research analyst at the Data Foundation, during the event.
“We continue to see CDOs citing limited budgets, lack of data-literate staff, as well as just a number of staff, and unclear CDO authorities as barriers to not only becoming a data-driven organization on that more individual level, but to the success of the CDO community at large and for being able to leverage innovative technology to support their missions,” she said.
The Federal CDO Council survey found that large agencies’ CDOs rated data culture, staff skills and limited data access as top obstacles in using data to support their agencies’ missions. Medium and small agencies also faced other obstacles like data governance challenges, lack of direct funding and tech barriers.