This interview is part of a series, Artificial to Advantage: Using AI to Advance Government Missions.

There are 58 chief AI officers across agencies large and small. Many wear multiple hats as the CAIO and the agency’s chief information officer or chief data officer.

No matter how many hats the chief AI officer wears, the foundation to make artificial intelligence work is the data.

Agencies must understand what it takes to make sure their data is ready to be used by AI tools.

To that end, the Commerce Department is reviewing responses to a request for information (RFI) from earlier this year seeking insights from industry experts, researchers, civil society organizations and other members of the public on the development of AI-ready open data assets and data dissemination standards.

At the same time, agencies are reviewing internal datasets and determining what needs to be done to make them ready for the AI tools.

Adita Karkera, the managing director of the government and public services practice at Deloitte Consulting, said the old adage of “garbage in, garbage out” has, maybe, never been truer than it is today.

“We really need to rethink and reprioritize the approach we’re taking for data management today because, of course, data is the fuel that’s driving AI applications today,” Karkera said on the discussion Data Readiness for AI and the Expanded CDAO Role. “The data that you’re going to provide as an input to these AI applications is going to drive how effective and trustworthy those AI applications are. As the volume of data around us is increasing, I think we just need to come up with the new and reprioritized methods of managing data, approaches that are going to help us reprioritize data quality, data security, data privacy and data trust.”

That means the CAIO, the CDO and other agency leaders must be at the forefront of managing data and prioritizing data policies so that the AI applications coming out are trustworthy and effective for citizens and employees to use.

Establishing a new kind of culture

Karkera said these leaders must create and maintain a data-driven culture and treat data as an asset for their organization.

“I think CDOs are really uniquely positioned to create that culture. We’ve seen a rise in the number of CDOs in the federal space today, which is really helping drive that data-driven culture,” she said. “Creating the data first mindset is definitely a deliberate effort for most CDOs that I’ve talked to. In fact, we have started to see that efforts like creating a data literacy program have become a key component of data strategies for chief data officers because we are no longer at a time where just knowing data is a choice. We have to be literate about data, and when I say data literate, it means that we have to have the mindset to be able to read, write and comprehend data and use it in ethical manners.”

At the same time for an agency to truly create that data-driven culture, Karkera said the CDO or chief AI and data officer (CDAO) can’t be on an island. Like any technology modernization program, executive level support as well as their day-to-day actions to use data to drive decisions are key to the organization’s success.

“They must lead by example so the rest of the workforce follows their lead and ask those same questions, makes those same efforts to ensure that the data that is in front of them is trustworthy,” she said. “While having the leaders lead by example is great, also having some sort of incentive program to say, ‘Okay, if you are showing your participation in data literacy programs or you’re going out and learning a new language or you’re learning a new tool, we are incentivizing it.’ In fact, Deloitte and the Data Foundation did a study in [March 2022] about data literacy in the government sector. We talked about executive sponsorship incentivizing programs as some of the key methods for how we can have successful data literacy deployed across the organization.”

The survey found the use of policies, grants and guidance would help employees or grantees incorporate expectations about data use in their relevant activities.

“Incentivizing the use of data may mean that certain traits or characteristics receive preference points in competitive grant programs, or that grantees are required to report on outcomes and also document methods for reporting such data,” the survey stated. “These incentives may even align with broader efforts to encourage evidence-informed policymaking in public sector organizations by recognizing and linking to program evaluation initiatives.”

The CDO’s role has expanded

Karkera said incentives can’t just be focused on connecting disparate data sets, but managing the quality and lifecycle of the data as well as creating the governance and rules around it.

A lot of those responsibilities fall to the CDO, whose roles and responsibilities have continued to evolve over the past five years.

Karkera said CDOs, especially now, are better positioned than ever to take the lead on creating, implementing and managing those data standards, policies and frameworks.

“Every step of the data lifecycle is so critical and CDOs are managing that today gracefully, while also making sure that the shared agendas of all other C-suite leaders are in perspective. I think about the CDO as being that glue that is bringing together other C-suite leaders and helping keep the focus on data, because data really is at the core of the enterprise’s or the agency’s mission goals,” she said. “I’ve started to see a lot of the data strategies that have been published by agencies now have data culture, data literacy programs, defining data and AI governance frameworks at the top of those strategy goals.”

Karkera added that agencies with more mature CDO roles and data governance and management strategies are more ready to begin using GenAI and other large language models in mission areas. She said those leaders are asking the right questions about where GenAI can solve mission problems in a trusted, secure and safe way.

“Everything starts with data, but we also have to ensure that as we are creating these AI solutions, we have the right frameworks in place to manage trust, bias, privacy and security,” Karkera said. “Are we really using data commensurate with the regulations that we collected that data for as we provide the data to the AI applications? Are we providing the right transparency, the right ethical boundaries around it? Because we are living at a time where humans and machines are working together, but what we don’t want to do is create solutions which are only machine based and have no human values at the center of it. I think I would tie this back to the CDO role because I think they’re so uniquely positioned to be embedded in this lifecycle and to ensure that they are providing the right ethical considerations as we are creating these solutions.”

For more in the series, Artificial to Advantage: Using AI to Advance Government Missions, click here.

Copyright © 2024 Federal News Network. All rights reserved. This website is not intended for users located within the European Economic Area.