Aydın Tiryaki (2026)

Introduction: The Futility of Bans and the “Cat-and-Mouse” Game

The integration of Artificial Intelligence (AI) in higher education has evolved beyond a temporary trend that can be ignored or curbed by simple prohibitions. The currently prevalent “catch and punish” methodology in the academic world has reached a technical dead end. Students can now process AI-generated texts through tools known as “AI Humanizers,” converting them into formats that bypass detectors and create the impression of being “manually written.”

The race between AI detectors and the tools that erase these digital footprints has turned into a “cat-and-mouse” game. Educational institutions cannot win this technological race through policing measures alone. The solution lies not in banning the technology, but in transparently integrating it into the educational process.

Step 1: Transparency and Declaration-Based Submission

In the proposed new model, the use of AI by students for assignments is not prohibited; on the contrary, the use of the most efficient tools is encouraged. However, this freedom is bound by the rule of “Absolute Transparency.”

• PDF Format and Declaration: Students are required to submit their assignments in PDF format. Within or attached to this document, it is mandatory to explicitly declare which AI tool was used and at which stage of the assignment (ideation, literature review, text editing, etc.).
• The Principle of Responsibility: The core message to the student is: “It is not essential who (or what) wrote the assignment, but who has mastery over the content.”

Step 2: Individual Verification (Appointment and Instant Quiz)

The most critical and innovative pillar of this system is the submission and evaluation phase. Instead of the classic “upload the file and wait” method, an active and dynamic process is proposed:

• Instant Quiz via Tools like NotebookLM: Today, tools such as Google NotebookLM can analyze an uploaded PDF file in seconds and generate questions specifically tailored to that document.
• The Appointment System: Assignment submissions are conducted through short (5-10 minute) scheduled appointments with the instructor or a teaching assistant.
• Process Workflow:
  1. The student arrives at the computer at the appointment time.
  2. The submitted PDF is uploaded to the system under supervision.
  3. The system instantly generates 3-5 “Mastery Questions” that require the student to defend their text.
  4. The student answers these questions on the spot.
• The Result: Even if the student had the assignment written by AI, if they have not read, understood, and internalized the text, they will fail to answer specific details about their own submission.

Step 3: Collective Analysis and Class-Wide Comparison

Once individual evaluations are complete, the system’s “Collective Assessment” capability comes into play. For example, when the same assignment topic is given to a class of 40 students, the 40 collected PDF files are uploaded into an AI-supported “Large Language Model” pool for cross-referencing. This stage yields the following outputs:

• Outliers and Originality Detection: Out of 40 assignments, 35 may have used similar AI models to provide similar “average” answers. The system instantly identifies the 5 “most original” assignments that deviate from this average, approach the topic from a different angle, or use rare sources. This is the most definitive way to distinguish between “standard AI output” and “effort-driven output.”
• Common Errors and Hallucination Detection: If a large portion of the class repeats the same erroneous information (AI hallucination), the system reports this as “This information appears in 12 assignments and is incorrect.” This proves that students used the information without verification.
• Similarity Clusters: Even if students are unaware of each other, structural similarities in the texts can reveal that they used identical prompts.
• Relative Grading: Originality ceases to be an abstract concept and transforms into concrete data measured by “how different the work is from the rest of the class.”

Grading Criteria and the “Sincerity Score”

The grading system is constructed on a triple hybrid structure:

1. Mastery (40%): The student’s performance in the instant quiz and oral defense (Viva Voce).
2. Originality and Collective Position (40%): How much the assignment distinguishes itself from the “herd” based on the collective analysis, and its accuracy.
3. Sincerity and Transparency (20%): The consistency between the student’s declaration and the technical analysis.

Conclusion: Campus Culture and AI Competition

This model will also transform the sociology of the university campus. Word will spread among students about which AI model is more efficient in which field, turning campuses into “Living AI Laboratories.”

• “Model X is better for engineering calculations.”
• “Model Y is more consistent for legal texts.”

Students will shift their focus from “how to cheat” to “which model helps create a more original synthesis.” This proposed system transforms universities from “detective agencies” into true academies where the most efficient collaboration of information, technology, and human intelligence is taught.

---

A Note on Methods and Tools: All observations, ideas, and solution proposals in this study are the author’s own. AI was utilized as an information source for researching and compiling relevant topics strictly based on the author’s inquiries, requests, and directions; additionally, it provided writing assistance during the drafting process. (The research-based compilation and English writing process of this text were supported by AI as a specialized assistant.)