Brady PieperChief information officers (CIO) and chief data officers (CDO) from 95% of U.S. states anticipate that the increased adoption of artificial intelligence (AI) will impact governments and businesses’ data management processes, a new report shows. Despite this, only 22% of states utilizing data management have an established AI-ready data quality program.
The National Association of State Chief Information Officers (NASCIO) report questioned CIOs and CDOs from 46 states to address concerns over data quality programs and their safeguards. The survey, completed in partnership with professional services firm Ernst & Young (EY), reflects states’ desperate need for established comprehensive, enterprise-wide data quality programs, which serve as a critical determinant in AI and generative artificial intelligence (GenAI) implementation.
“This report demonstrates that there is a clear imperative for state leaders to implement data quality programs on the path to successful AI integration,” NASCIO Executive Director Doug Robinson says in a press release. “Creating a data-centric culture within state CIO offices will help leaders identify patterns that will drive innovation and improve the quality of citizen services.”
RELATED: States taking on challenge of Artificial Intelligence legislation
According to the report, effective data quality programs are essential for state governments’ data management and identifying the best return on investment for data quality efforts. For data quality programs to be effective in supporting AI and GenAI applications, they must achieve advanced maturity with pristine datasets, automated data validation and consistent data standards. Additionally, robust key performance indicators and dynamic data quality rules are essential to adapt to changing needs, ensuring that AI outputs support informed policymaking and positive outcomes.
The NASCIO and EY report outlines seven key considerations for achieving AI-readiness in data management, reflecting the growing recognition that high-quality data is critical to leveraging AI and GenAI effectively. These seven considerations include:
- Develop and execute AI-ready data strategies that enable rapid AI and GenAI solutions.
- Reorganize and integrate organizational knowledge, traditionally stored outside of data management, into key data quality programs.
- Employ data scientists and AI/machine learning engineers to utilize novel data sources and optimize datasets for AI and GenAI solutions.
- Provide a single source of truth for core datasets, reducing inaccuracies in AI responses and improving AI trust.
- Fix and improve inaccurate, poorly maintained data and improper access controls to reduce the risk of incorrect AI and GenAI solutions.
- Ensure the quality of AI and GenAI solutions through curated and validated data.
- “Thoughtfully” curated data that enables rapid prototyping and expansion as AI and GenAI solutions scale.
Despite less than a quarter of organizations utilizing data quality programs, 89% of respondents rate their organization’s data quality as important, very important or critically important to operations according to the survey. NASCIO and EY state that this lack of investment in data management causes a “cyclical pattern” of manual re-cleansing that leaves organizations constantly repairing old datasets instead of creating ongoing monitoring mechanisms and designated data stewards that proactively address data concerns.
According to the report, 72% of respondents rank their organization’s data quality maturity as aware or reactive, reflecting an organizational lack of understanding of data quality principles and deficiency in fostering data quality initiatives.
Legislative and organizational policy, current organizational investments and trained workforce availability highlight some of the hindrances to AI and GenAI implementation outlined by the report.
In regard to policy, states differ in their opinion of data collection and ownership. For example, Utah is facing challenges in prioritizing important data delivery policies, including data privacy, data sharing and cross-organization data standardization and collaboration. In Oregon, state organizations pull data from several out-of-organization partners, enforcing the state’s desire for the collaborative development of policies and standards between the state and its partners.
RELATED: Harnessing AI: How federal agencies are innovating for the future
Organizational investment into data quality initiatives does not currently align with the need, according to the report. It shows that only one respondent of 48 ranked their organization’s funding as highly aligned with their data quality importance.
Regarding workforce availability, NASCIO and EY write that there is a “notable gap” in the availability of skilled personnel required to build an operational framework that maintains data quality programs. NASCIO recommends workforce and resource pooling to improve this concern, which will also further the interoperability of data standards.
NACSIO and EY conclude their report by providing recommendations for organizations to prepare for AI and GenAI implementation.
One recommendation is for organizations to implement comprehensive data governance to ensure strategic alignment. This involves establishing a data governance framework with oversight and formalized standards, assessing data assets based on their impact and setting data quality requirements to align with business needs.
Additionally, organizations should implement data quality standards at dataset creation, select effective tools, institute robust access and security policies and maintain detailed metadata. The report also states that investing in data literacy and engineering capabilities will further support the development of reliable and actionable data for AI and GenAI implementation.
“AI technology is set to transform the way state agencies operate and innovate, but success is highly dependent on the quality of the data,” EY U.S. State and Local Technology Leader Chris Estes says in a press release. “Prioritizing data quality management and funding is critical to realize the full value of state government AI and technology modernization efforts.”
Photo by Markus Spiske on Unsplash
