Professors Are Exhausted by AI. Here Is What Faculty Development Should Do About It.

Some of the most dedicated teachers in America are quietly walking away.

The Chronicle of Higher Education has reported on professors retiring early with AI as the last straw, including an English professor at Southern New Hampshire University who said AI was part of her decision to leave. She is not alone. Across the country, faculty describe redesigning courses on nights and weekends, holding hallway conversations in place of institutional guidance and carrying the emotional weight of accusing students they want to trust.

As chair of a humanities department at a Chicago community college, I see this up close. The question faculty bring me is not "How do I use ChatGPT?" It is closer to: "What is my job now?"

That is a philosophical question. It deserves better than a one-hour webinar.

The Numbers Behind the Exhaustion

The 2026 national faculty survey by the American Association of Colleges and Universities (AAC&U) and Elon University's Imagining the Digital Future Center gives the scale.

• 68 percent of faculty said their institutions have not prepared them to use generative AI in teaching and mentoring.
• 59 percent said students are not being prepared for an AI future.
• Nearly 80 percent expect AI to change the typical teaching model in their department within five years.
• About one in four faculty never use generative AI at all, and a third do not use it in teaching.

Read those together and the picture is clear. Faculty know a transformation is coming. Most feel unequipped for it. And many are understandably skeptical of the tool itself.

The Chronicle's reporting adds the human texture. Over 80 percent of the instructors who responded to its reader survey had tried to AI-proof assignments, and 65 percent had caught students using AI to cheat. The work is falling heavily on contingent faculty and graduate instructors who teach the large introductory courses, often with the least support.

Why Most AI Training for Faculty Misses

Much of what passes for AI faculty development falls into two traps.

The tools trap. Sessions that demonstrate features, prompts and platforms. Useful for some, but they answer a question most faculty are not asking. A professor worried about whether her essays still measure anything does not need a tour of a new interface.

The policy trap. Sessions that review the institution's AI policy and the academic integrity process. Necessary, but defensive. They position faculty as enforcers rather than designers.

Both miss the deeper need. Faculty need a way to rethink what learning looks like when a machine can produce the artifacts we used to grade. That is a question about purpose, evidence and judgment, which is to say, a philosophical question about teaching.

What Good AI Faculty Development Looks Like

Drawing on the research and on what is working in classrooms, effective programs share five features.

1. Start With Purpose, Not Platforms

Begin with the question every discipline must answer: what kind of thinking do we want our students to be able to do without help? Once a department names that, decisions about AI become much clearer. Some skills should be practiced with AI. Some must be practiced without it.

2. Redesign Assessment Together

Cornell researcher René Kizilcec and colleagues, writing in Science in 2026 after studying more than 95,000 students, concluded that "assessment reform is necessary and urgent." They recommend clearer rules on acceptable AI use, some return to controlled in-person assessment and assignments in which AI use demonstrates professional skill. Faculty need structured time to rebuild real assignments with colleagues, not just a list of tips.

3. Make Room for Disagreement

Faculty do not agree about AI, and they should not be forced to pretend they do. In the AAC&U survey, faculty were split nearly evenly on whether an AI-generated outline counts as cheating. Good development creates space for honest disagreement and helps each instructor state a clear, defensible position in the syllabus.

4. Keep the Human Side of Teaching in View

The relationship between teacher and student is under strain when every assignment invites suspicion. Faculty development should address trust directly: how to talk with students about AI, how to respond to suspected misuse without becoming a prosecutor and how to keep the classroom a place of formation rather than surveillance.

5. Support the Work Over Time

A single workshop cannot carry a redesign. The most promising models pair an initial workshop with a semester-long faculty learning circle, where instructors test redesigned assignments, share results and adjust. Adjunct-friendly formats, with pay or professional development credit, matter, because adjuncts teach so many of the courses most affected.

The Philosophical Core: Teaching Judgment

Here is the conviction that shapes my own faculty work. AI is excellent at producing answers. It is not responsible for them. Students will enter a world in which the most valuable human skill is not generating content but judging it: recognizing a flawed assumption, testing a persuasive claim, deciding what matters and taking responsibility for the decision.

The philosopher and computer scientist Brian Cantwell Smith calls this the difference between reckoning, the calculating power machines excel at, and judgment, the capacity to be responsible to the world as it really is. Every discipline can teach judgment. Most already do. Faculty development should help them see that this is now the center of their work, not the margin.

Seen that way, AI does not make professors obsolete. It makes the best parts of their job more important than ever.

Questions for Deans and Teaching Centers

• Have our departments named the thinking skills students must be able to perform without AI?
• Do faculty have paid, structured time to redesign assessments together?
• Do our adjunct and graduate instructors have access to the same support as full-time faculty?
• Are we training faculty as designers of learning, or only as enforcers of policy?
• Will our support last a semester, or end when the workshop does?

The professors retiring early are telling us something. The answer is not to wait for the technology to settle down. It is to give the people at the heart of education the time, tools and vision to lead through it.

Sources and Further Reading

• AAC&U and Elon University Imagining the Digital Future Center, national faculty survey on generative AI (2026). Chronicle of Higher Education summary.
• "These Professors Are Retiring Early. AI Was the Last Straw," Chronicle of Higher Education (2026).
• "AI Is Overhauling the Way Professors Teach," Chronicle of Higher Education (2026).
• René F. Kizilcec, Igor Chirikov and Ivan Smirnov, "Generative AI Use and Misuse Call for Assessment Reform in Higher Education," Science (2026).
• Brian Cantwell Smith, The Promise of Artificial Intelligence: Reckoning and Judgment (MIT Press, 2019).

Planning a spring development day or faculty learning series? Dr. Sweis offers workshops including "Teaching Thinking When AI Can Write the Essay" and "What AI Can't Do: Judgment, Ethics and the Human Side of Learning." Bring a workshop to your campus.