A query using Artificial Intelligence (AI) ChatGPT surfaced several items by Jason Dom, a Director at the American Public University System (APUS), who sees a legitimate role for AI in higher education. However, its use should be transparent, course-specific, and governed by faculty and institutional rules. For example:
The central idea: Don’t simply ban AI
APUS’s published guidance explicitly says the university embraces generative AI as a learning tool and encourages students to understand both its benefits and risks. The university’s current policy permits or encourages AI use in some courses and assignments when the instructor or assignment directions say so.
The APUS model says AI can be part of the educational process, but the student remains responsible for demonstrating learning and following the rules of the particular course in which they are enrolled. Students should use AI “to learn—not to avoid learning.”
Several suggested key priorities.
A. Faculty control
The professor—not the AI—remains responsible for determining how AI fits into the course.
ibl.ai is an AI Faculty Agent explicitly describes its role as supporting instructors while keeping the instructor “firmly in control of all academic and pedagogical decisions.”
B. Guardrails rather than unrestricted generation
An educational AI tool should be configured to say, “I can help you understand this concept, but I won’t write the answer to your graded assignment.”
C. Institutional context when using AI tools
The academic institution should direct students to specific campus resources. For example, ibl.ai describes its architecture as allowing institutions to connect AI agents to systems such as LMSs and student-information systems while retaining control of data and deployment.
D. What all this means for students
The APUS approach described by Jason Dom effectively creates three different categories of AI use.
APUS permitted learning assistance by AI tools
Various permitted uses of AI include:
- brainstorming;
- research assistance;
- explanations of difficult concepts;
- study guides;
- flashcards;
- summaries;
- editing assistance, subject to course rules.
AI assisted work requiring disclosure
APUS says AI generated material that appears in a student's submission must be appropriately cited or attributed, and students may be required to explain how and to what extent they used AI.
“What did you use it for, and what intellectual work did you personally perform?”
Prohibited substitution
APUS currently identifies inappropriate use of AI, among other things:
- having generative AI write an assignment or discussion post;
- failing to verify AI-generated information;
- failing to provide appropriate attribution;
- using AI during tests/exams/quizzes without permission.
So the student’s fundamental obligation remains: The work submitted for academic credit must actually demonstrate the student’s learning.
What does all this mean for faculty
Faculty aren’t simply being asked to detect AI-generated papers. They’re being encouraged to design the learning environment around AI’s existence.
ibl.ai’s Faculty Agent, for example, is designed to help faculty:
- develop syllabi;
- create assignments;
- develop rubrics;
- generate examination questions;
- prepare lectures;
- assist with grading and feedback;
- identify assignments that are easily completed by AI;
- redesign those assignments when appropriate.
The professor remains the decision-maker.
The professor decides which criteria actually belong in the course.
The biggest implication for academic integrity
This is probably the most interesting aspect of the APUS and ibl.ai models.
Traditional academic-integrity systems largely ask: “Did the student cheat?”
But AI forces universities to ask a more complicated question: “What constitutes authentic student work when AI is an ordinary part of the student’s intellectual environment?”
APUS must distinguish between assistance and substitution.
AI-assisted work requiring disclosure
APUS says AI-generated material that appears in a submission must be appropriately cited or attributed, and students may be required to explain how and to what extent they used AI. That creates a principle of AI transparency.
The question is “Did you use AI?” and “What did you use it for, and what intellectual work did you personally perform?”
Prohibited substitution
APUS currently identifies as inappropriate, among other things:
- having GenAI write an assignment or discussion post;
- failing to verify AI-generated information;
- failing to provide appropriate attribution;
- using AI during tests/exams/quizzes without permission.
So the student’s fundamental obligation remains: The work submitted for academic credit must actually demonstrate the student’s learning.
What this means for faculty
This model puts considerably more responsibility on instructors. Faculty aren’t simply being asked to detect AI-generated papers. They’re now being encouraged to design the learning environment around AI’s existence. That is a major conceptual shift.
ibl.ai and its Faculty Agent, for example, is designed to help faculty:
- develop syllabi;
- create assignments;
- develop rubrics;
- generate examination questions;
- prepare lectures;
- assist with grading and feedback;
- identify assignments that are easily completed by AI;
- redesign those assignments when appropriate.
But the system explicitly positions AI outputs as drafts requiring faculty review, rather than autonomous academic decisions. That is important.
The professor remains the decision-maker.
The professor decides which criteria actually belong in the course.
The professor decides whether and how to redesign it.
The instructor retains responsibility for the final judgment.
The Jason Dom/s APUS + ibl.ai approach can be understood as a move from AI prohibition and detection toward controlled educational integration.
For students, AI becomes a legitimate learning resource—but not a substitute for their own intellectual work.
For faculty, AI becomes a teaching and productivity tool—but the instructor remains responsible for the pedagogy, assessment, and final academic judgment.
For academic integrity, the key distinction becomes authorized assistance versus unauthorized substitution.
For institutional policy, AI governance expands beyond plagiarism to include assessment design, disclosure, privacy, data ownership, model selection, access controls, faculty autonomy, and accountability.
To close, many of the lessons and recommendations presented by Jason Dom on the use of AI tools in academia also can be applied to other industries.
