The report, 2025 EDUCAUSE Horizon Action Plan: Building Skills and Literacy for Teaching with Generative AI, is the latest in EDUCAUSE’s long-running Horizon series and outlines how educators and administrators can build skills and literacy to teach with generative AI now and in the future.
“Nobody can predict the future, but we can plan for the future,” said Jenay Robert, a senior researcher at EDUCAUSE who authored the report. “That is the real magic of the Horizon Report, and the Horizon Action Plan in particular, is being able to envision where we want to go, being intentional about planning for that preferred future, and then walking through the steps that will get us there.”
SHAPING THE IDEAL
In the preferred future described in the report, higher education renews public trust through transparency efforts, bridges digital divides by sharing resources and literacy initiatives, improves teaching through generative AI, and enhances collaboration through new governance frameworks. To move toward that vision, the plan outlines actions across four levels: individual, departmental, institutional and multi-institutional.
At the individual level, EDUCAUSE urges educators to spend time experimenting with generative AI and researching ethics, privacy and intellectual property. Robert said the process starts with professional curiosity, encouraging staff to try out AI tools in low-stakes ways, such as drafting course announcements or generating quiz questions, before integrating them into assessments or research.
At the unit or department level, institutions are encouraged to integrate AI into curricula, create flexible teaching guidelines and offer ongoing professional development. The report recommends using generative AI to speed up administrative tasks while evaluating its role in assessment and feedback. Departments can also start by reviewing learning outcomes through the lens of AI literacy, asking how students might use, critique or improve AI tools as part of their work in their discipline of choice. EDUCAUSE recommends embedding AI-related concepts in both general education and advanced coursework.
For larger institutional initiatives, EDUCAUSE calls for clear ethical frameworks, institution-wide discussions on academic integrity and working groups of IT staff, instructional designers and accessibility professionals to assess best practices.
At the multi-institutional level, the report stresses the value of collaboration, including sharing case studies, developing common guidelines and aligning learning outcomes across institutions or even throughout the world.
Robert said the California State University (CSU) system is a good example of cross-institutional collaboration, bringing the same AI tools and training to all of its campuses while acknowledging that each site will bring its own resources and goals. CSU’s AI in July series, open to all campuses, is one model of how shared investment can close gaps between institutions, Robert said.
CHALLENGES
Still, the path forward is complex. Disciplinary differences make it difficult to set universal standards.
“What might be considered perfectly fine and appropriate in a chemistry lab — maybe writing an email to my co-PI [principal investigator], or trying to refine my synthetic description of my synthesis that I did for some chemical — this all seems perfectly reasonable, and then when you move into something like education or any social science, writing becomes a bit different,” Robert said.
To navigate these differences, Robert said institutions should establish broad ethical principles while leaving space for departments to decide how to apply them. In practice, that means faculty groups or curriculum committees can adapt guidelines based on disciplinary norms rather than adopting one-size-fits-all rules.
Smaller institutions also tend to lack the funding or expertise to keep pace, according to EDUCAUSE. Robert said those campuses should look first to consortia, statewide systems or associations for shared resources and professional development.
Evaluating the return on investment for ed-tech tools remains difficult as well, Robert said.
“It’s not as clear cut as, ‘This tool saves us dollars and cents,’” she said. “Some of those things that are really important to us in higher education, specifically around student experience — Are students more engaged? Do students feel a greater sense of belonging? — or even measuring some specific learning outcomes, can be really challenging.”
EDUCAUSE suggests program evaluation methods, like reviewing qualitative faculty feedback and student performance data, may be more useful than traditional cost-based ROI models.
As institutions consider their future, Robert encourages educators to be "aspirational."
"We know that we can't craft the exact future that we want to see," she said. "But if we're aiming at nothing, that's where we're going to land: nothing."
