AI Won’t Fix Broken Systems

Artificial intelligence is moving through higher education with remarkable speed. Institutions are evaluating tools, forming committees, creating policies, and looking for ways to increase efficiency.

That energy can be productive. AI can help institutions analyze information, reduce repetitive work, improve access to knowledge, and create more responsive experiences.

But AI cannot compensate for a process that should not exist, a policy no one can explain, or a system designed around institutional convenience instead of human needs.

When technology is added to a broken system, it often allows the broken system to operate faster and at a greater scale.

Technology amplifies the system around it.

Institutions often approach technology as if it can be separated from the environment in which it operates. We choose a platform, configure it, train employees, and expect better outcomes.

But every technology inherits the assumptions, policies, workflows, and power structures surrounding it. If those structures are confusing or inequitable, technology can reproduce that confusion and inequity more efficiently.

Imagine an institution using AI to answer students’ questions about a complicated administrative process. The AI may make the instructions easier to find, but it does not necessarily make the process itself easier to complete.

We may have improved the explanation without improving the experience.

We keep treating systems problems as people problems.

When students struggle to navigate an institution, we often describe the problem in individual terms. Students need to be more prepared, more engaged, more proactive, or more persistent.

Employees receive similar messages. They need to communicate more effectively, collaborate more often, learn another platform, or work around another process.

Sometimes individual support is necessary. But when the same problem affects many people repeatedly, we should question the system—not simply the people experiencing it.

• If students repeatedly miss a requirement, the requirement may not be communicated clearly.
• If employees maintain separate spreadsheets to complete routine work, the official system may not support the actual workflow.
• If every successful outcome requires someone to intervene, the institution may be relying on heroics as an operating model.
• If students must tell the same story to five offices, the institution may have created an integration problem and assigned it to the student.

Humans have become the integration layer.

In many institutions, systems do not communicate well with one another. Departments operate with different processes, definitions, timelines, and sources of information.

People compensate for these gaps. Employees translate between systems, manually transfer information, maintain shadow records, and explain institutional contradictions to students.

Students perform integration work too. They piece together emails, websites, portals, policies, forms, and advice from different offices to understand what they are supposed to do.

AI might help people complete some of this invisible work. But before automating it, leaders should ask why the work exists and whether it creates value.

Redesign before you automate.

The goal is not to delay every AI initiative until an institution achieves perfect alignment. No institution has perfect systems, and meaningful change often requires experimentation.

The goal is to resist automating a process simply because it already exists.

Map the experience. Talk to the people closest to the work. Identify unnecessary handoffs and unclear decisions. Examine where people create workarounds. Simplify what can be simplified.

Then determine where AI can strengthen the redesigned experience.

AI strategy is organizational strategy.

AI decisions are not only technology decisions. They are decisions about institutional priorities, employee roles, student expectations, information access, risk, trust, and the kind of experience an institution wants to create.

That means AI strategy cannot belong exclusively to the technology office. It requires academic leaders, student affairs professionals, operational teams, faculty, staff, students, and technology experts to examine the system together.

The institutions that benefit most from AI will not necessarily be the ones that adopt the greatest number of tools. They will be the ones that understand what they are trying to improve—and have the courage to redesign the systems that technology will amplify.