Too often, institutions are expanding access without building the systems needed to support meaningful, equitable, and sustainable use. The result is a growing gap between what institutions hope to achieve and what is actually happening in practice.
We conducted focus groups and interviews with faculty, administrators, IT leaders, and students, alongside a national survey with 120 respondents from institutions of higher education across the U.S. This work is rooted in the lived experiences of the people doing the work. Rather than proposing another framework, our recent report examines how institutional conditions shape what actually happens on the ground. We provide five findings from the study with corresponding recommendations to surface where systems are breaking down and where they can be strengthened to support coordinated, scalable AI integration.
Across institutions, AI adoption is largely driven by individual faculty initiative rather than institutional strategy. Faculty described having the autonomy to experiment, but this often leads to a haphazard approach across campus. Less than 30% of respondents agreed that technology decisions and policies are clearly communicated across their institution.
Institutions have moved quickly to provide AI tools. Over 70% of survey respondents reported that students have access to paid AI tools. But policies, learning goals, and instructional guidance have not kept pace. Faculty and students are navigating mixed messages. Without shared institutional direction, one instructor may encourage using AI for brainstorming while another prohibits it entirely, leaving students confused about expectations.
Professional development on AI is widely available, but it is often misaligned with faculty time, workload, and incentives. Alarmingly, 60% of faculty reported they are not given compensated time to integrate new technologies. Workshops are often too long, resulting in “cognitive overload,” or are disconnected from classroom needs. Instead, what works are shorter, applied sessions.
Nearly all institutions recognize the importance of preparing students for an AI-driven workforce. But implementation is uneven. AI learning opportunities are often concentrated in STEM programs, while students in other fields have limited exposure. As one administrator noted, “STEM students are getting all the love… but no one is really looking at scale at the non-STEM students”. Consequently, about half of faculty reported lacking resources to understand how AI connects to workforce applications.
Infrastructure, funding, and staffing limitations make equitable implementation difficult. Institutions are piloting promising initiatives, often with grant funding, but many struggle to scale them without sustained resources once the funding ends. Faculty even reported abandoning useful tools if they weren’t centrally provided, knowing that only students who could afford private subscriptions would benefit.
Across all five areas, a clear pattern emerges. Institutions are moving quickly, but not always coherently. AI integration is not just a technology challenge. It is a systems challenge. When governance includes the right voices, when guidance is clear, when professional learning fits faculty work, when efforts are coordinated across disciplines, and when equity is resourced, AI can move from isolated innovation to institutional transformation.
Higher education is at a critical moment. The question is no longer whether to adopt AI. It is whether institutions will build the systems needed to do so thoughtfully, equitably, and at scale.
