If AI Can Teach, What Becomes the Role of Faculty?
Imagine a student trying to understand a difficult reading late at night. They ask an AI tool to explain a confusing passage. It offers a simpler explanation, gives an example, and answers a follow-up question. The student gets unstuck and keeps going.
That can be a useful moment in learning. It also raises a bigger question: If AI can teach, what becomes the role of faculty?
The answer depends on what we mean by teaching. If teaching means delivering explanations and information, AI can already do parts of it. It can rephrase an idea, generate examples, offer practice questions, and respond to students when an instructor isn’t available.
But teaching has never been only the delivery of information.
A faculty member makes choices about what matters in a subject, what students are ready to learn, and what they should be able to do with that knowledge. They notice when a question reveals a deeper misunderstanding. They decide when to offer guidance, when to let students work through difficulty, and when the difficulty has become a barrier to learning.
Those choices require judgment about the student, the subject, and the moment.
Research on human tutoring helps explain why. In a study of tutoring interactions, Chi and colleagues found that students learned effectively even when tutors were discouraged from supplying explanations and feedback and instead prompted students to do more of the thinking. The researchers connected that learning to students’ active construction of ideas and participation in the exchange. The value of tutoring, in other words, did not come only from the tutor having answers. It also came from how the tutor engaged the learner. Wiley Online Library
AI may provide a helpful explanation without knowing whether it is what a particular student needs. It may respond to the question asked without recognizing the question the student is still trying to formulate. And a student may find an answer clear without being able to explain, apply, or evaluate it. A response can sound like teaching without producing learning.
A 2025 systematic review of AI tutoring systems found examples of systems that adapt to students’ progress and support learning. The review focused on K–12 studies, so its findings should not be treated as direct evidence about higher education. Still, it illustrates both the potential of AI tutoring and the importance of asking what learning outcomes these systems support and what they leave for educators and students to do. npj Science of Learning
An emerging systematic review of teacher intervention in AI-supported instruction makes a related point. Reviewing 29 studies, the researchers describe a process in which teachers interpret AI-generated information, use professional judgment, and decide how to respond in context. The review also cautions that its findings come from K–12 settings. It offers a useful question for higher education, though: AI-generated feedback or recommendations do not become meaningful instruction on their own. Someone has to consider what they mean for learners and what should happen next. Springer Nature Link
Students need more than answers. They need opportunities to develop the knowledge and judgment to ask better questions, evaluate answers, and use what they learn in new situations. An AI tutor might help a student get started. Faculty can help students examine why one answer makes sense, what evidence supports it, and where it might not apply.
The aim isn’t to withhold help. It’s to make sure that help supports the student’s learning instead of doing the important thinking in the student’s place.
That distinction changes how we think about the faculty role. Faculty do not need to compete with AI at producing explanations or information. They can design experiences in which students use information, make decisions, test ideas, and reflect on what they understand. They can connect course material to a discipline’s methods and responsibilities, and respond to students as people learning within a shared community.
AI may take on more instructional tasks. As it does, faculty decisions about the purpose of a course, the design of its learning experiences, and the support students receive become all the more visible.
If an AI system explains a concept, we can still ask whether students can use it. If an agent guides practice, educators still decide what kind of practice matters. If AI responds to student questions, institutions and faculty must decide when a human conversation is needed.
AI may teach parts of a course. Faculty remain responsible for helping determine what students are learning, how they will learn it, and what it means to become capable in a field.
That leads to the next question in this series: If AI can generate responses to assignments and perhaps complete them, what should assessment be designed to show?
Continuing the Conversation
Series 1: AI Is Exposing Existing Problems ✓ Completed
Series 2: What We Do About It ✓ Completed
Series 3: Cultivating Human Thinking in an AI World ✓ Completed
Series 4: Learning Alongside AI ✓ Completed
Series 5: Preparing Students for an AI World ✓ Completed
Series 6: The Purpose of Education in an AI World
Current Post (1 of 8): If AI Can Teach, What Becomes the Role of Faculty?
Next Up: If AI Can Complete Assessments, What Is Assessment For?