Data Literacy Becomes Essential for AI ROI in Higher Education
Morgan State University's dean can pull live enrollment data in minutes. At most colleges, the same request takes three weeks. The difference isn't the technology - it's how the historically Black university prioritizes data literacy among its staff.
Timothy Summers, vice president of IT and CIO at Morgan State, is building the institution's AI strategy around employees who can interpret, question and act on data. Without that capacity, universities pay for capabilities their people cannot use.
Morgan State built Obsidian, a sovereign AI system created by its own engineers that learns from the institution and records every interaction for transparency. The university is the only HBCU on Google's Research Technology Leaders Advisory Board, which includes Stanford, Yale and Johns Hopkins.
The Data Literacy Gap
Faculty and staff must actually understand what data tells them for any of this to work. EDUCAUSE's 2026 Top 10 identifies data literacy as one of the most underdeveloped skills in higher education.
A National Skills Coalition analysis of 43 million job postings found that 92% of U.S. jobs require digital literacy skills. Yet nearly a third of the American workforce has little to no digital literacy. In an AI environment, that gap becomes costly.
Data literacy today means more than spreadsheets and charts. "This is about the institutional capacity to engage critically with data, to interpret it, to question it," Summers said. Staff need to spot where data is incomplete or biased and make responsible decisions in AI-driven systems.
Chris Hein, field CTO at Google, defines data literacy as "the ability to make sense of all of the volumes of data that somebody might have access to" and "being able to identify what is the signal and the noise in all of the information that's coming to you."
Why Data Literacy Matters for AI Readiness
With AI in higher education, data no longer supports just one person's decision. Matt Jubelirer, general manager of education marketing at Microsoft, explained: "It becomes foundational context for the next person, the next workflow."
Governance matters at the infrastructure level too. How data is stored, protected and structured determines what AI can actually do. Data protection and storage solutions help institutions establish the clean, governed data foundation that AI requires, including securing sensitive research and student data against compliance risks.
"Data literacy now includes understanding how data is created, governed and used responsibly across the institution," Jubelirer said. "Something as simple as a confidentiality label or storage location can directly impact how data can be used across teaching, research and operations."
Summers put it plainly: "An institution that can't read its own data can't govern its own AI."
Measuring Staff Data Literacy
Self-assessments alone don't measure institutional data literacy. Summers emphasized pairing them with behavioral indicators: Can a staff member interpret a retention dashboard responsibly? Do they understand what the numbers don't say? Can they validate an AI recommendation against institutional context before acting?
Jubelirer recommends that institutions map key roles across campus and identify which AI-enabled workflows matter most to each one. The question becomes whether those workflows are powered by trusted, high-quality, appropriately governed data.
Microcredentialing and badging can validate literacy levels. Staff earn credentials by completing tasks that require data competency and demonstrating understanding of the data itself.
Building a Structured Framework
Morgan State developed an AI literacy framework with three progressive stages: I can perform, I can lead and I can decide. Purdue University and other institutions are taking similar structured approaches using Google's Gemini for Education to establish tiers around data literacy pathways.
"With a solid framework, you've got a little bit less of a Wild West scenario," Hein said.
Building that structure requires organizational infrastructure to sustain it. A growing number of institutions are standing up AI centers of excellence to provide governance structure, cross-functional alignment and literacy programming. Without that coordination layer, even the best frameworks tend to stay siloed in IT rather than scaling across campus.
Self-Service Analytics and Governance
Morgan State's framework supports broader institutional access to trusted data. Self-service analytics - pulling enrollment reports on demand - happens "when you build access and literacy simultaneously," Summers said.
That accessibility requires strong governance underneath. "Data must be connected, trusted and aligned to privacy and compliance requirements," Jubelirer said. Institutions must ensure the shared data layer maintains quality and security standards before broadening access.
Institutions that don't build access intentionally tend to attract shadow AI tools that bypass governance and introduce the exact data privacy and compliance risks they're trying to avoid.
Measuring ROI
Schools can measure data literacy's impact by looking at overall AI transformation across the institution. Are faculty spending more time supporting students and less time on administrative work? Are advisers identifying and acting on risk earlier?
The return on data literacy "doesn't show up in a training dashboard. It shows up where the return on AI shows up," Summers said.
"It shows up in a student who stays enrolled because an adviser caught a retention signal three weeks before it became a dropout. It shows up in a financial aid office that identified a processing bottleneck before it created 400 calls in a single week."
By measuring those outcomes, schools can demonstrate how data analysis literacy drives ROI around AI for education.
Your membership also unlocks:
