Özyeğin University AI User Guide

Introduction
This document outlines the principles to be observed regarding the use of AI technologies in all academic and administrative processes of the University, including but not limited to educational activities, research and development studies, evaluation and assessment processes, and student support services.

The first section of the document provides functional definitions of frequently used concepts related to artificial intelligence. The second section sets out essential principles, which are also included in the University's AI Policy. The third section specifies the principles that students, academic and adminsitrative staff should consider when using AI.

1. Definitions
The AI-related concepts used in this document are defined below based on internationally recognized sources. As artificial intelligence technologies continue to evolve, these definitions may be reviewed and updated when necessary.

Algorithmic Bias:  The systematic production of unfair outcomes by artificial intelligence systems toward certain individuals, groups, or situations due to imbalances in training data or limitations in algorithmic design.

Academic Integrity:  A fundamental academic principle requiring honesty, trust, fairness, respect, responsibility, courage, transparency, and ethical conduct in all academic activities.

Artificial Intelligence (AI): Methods and technologies that enable computer systems to perform tasks requiring human intelligence, such as learning, reasoning, pattern recognition, decision-making, or understanding natural language.

AI Agent (Agentic AI): An AI system capable of planning towards a specific goal, performing multi-step tasks, and operating autonomously or semi-autonomously using various digital tools.  

AI Literacy: The ability to understand, critically evaluate, and responsibly use AI systems while recognizing their capabilities, limitations, ethical implications, and societal impacts.
 
AI-Powered Application: A digital system or software that uses artificial intelligence components in some of its functions and provides services to users directly or indirectly.  

AI Provider: A natural or legal person who develops, markets, or offers an artificial intelligence system under their own name.  

AI System: A machine-based system that generates outputs, such as predictions, content, recommendations, or decisions, from the inputs it receives, in accordance with specific objectives, and that can influence physical or virtual environments through those outputs.
 
AI Tool: Software, applications, or digital services developed to perform specific tasks using artificial intelligence technologies such as machine learning, natural language processing, and image recognition.  

AI User: An individual or organization that uses an artificial intelligence system in their activities or operations.  

Chatbots: AI-based dialogue systems designed to provide information, give directions, or perform specific tasks by interacting with users via text or voice.
 
Cyber Vulnerability: A security flaw that weakens the security of information systems, networks, software, or digital infrastructure and, if exploited by malicious actors, could lead to unauthorized access to systems, data loss, data manipulation, or service disruption.

Cyberbullying: The intentional and repeated use of digital communication tools to inflict harm on individuals, including threatening, humiliating, defaming, harassing, or excluding.  

Data Minimization: The data protection principle that requires that only the minimum amount of personal data necessary to achieve a specified purpose be collected, stored, and used.  

Data Masking: A data protection technique that ensures sensitive or personal data is modified in a way that preserves its functional use, while making it unreadable by unauthorized individuals or impossible to link to the data subject.  

Deep Learning (DL): A machine learning approach that uses multi-layered neural networks to learn complex patterns in large datasets.  

Deepfake: The creation of fake visual, video or audio content that closely mimics a person's face, voice, or behavior using deep learning techniques.

Digital Learning Assistant: An AI-based digital assistant that helps students with their learning processes by assisting them with tasks such as accessing course materials, suggesting content, answering questions, or academic planning.   

Digital Teaching Assistant: Software, applications, or systems that use AI or other digital technologies to support teaching activities such as lesson preparation, content creation, evaluation and assessment, feedback, and managing learning processes.

General Data Protection Regulation (GDPR): A data protection regulation adopted by the European Parliament and the Council and entered into force on May 25, 2018, governing the processing of personal data in the European Union and aiming to strengthen individuals’ rights regarding their personal data.

Generative Artificial Intelligence (GenAI): AI models trained to generate new content such as text, images, audio, video, or software code.

Fine-Tuning: The process of re-adapting a pre-trained artificial intelligence model to a specific task or domain by using additional data.

Hallucination: The generation of false, fabricated, or unverifiable information by an artificial intelligence system that is presented as factual or accurate.

Large Language Model (LLM): An artificial intelligence model trained on large-scale text data, capable of understanding, interpreting, and generating natural language.

Machine Learning (ML): A set of methods that enables computers to learn patterns from data without being explicitly programmed, and to improve their performance on tasks through this learning process.

Model Training: The process through which an artificial intelligence model learns from data to perform specific tasks.

Natural Language Processing (NLP): A field of artificial intelligence that employs methods and techniques to enable computers to understand, interpret, analyze, or generate human language.

Peer Bullying: Intentional and repeated behavior directed at an individual by peers, involving a real or perceived power imbalance and causing physical, verbal, psychological, or social harm.

Personal Data: Any information relating to an identified or identifiable natural person. Examples include name, contact information, academic performance data, IP address, voice recordings, or facial images.

Prompt:​ A text, question, command, or instruction provided by a user to an artificial intelligence model to generate a specific output.

Personal Data Protection Law (KVKK): The principal data protection law of the Republic of Türkiye, which entered into force on April 7, 2016, and sets out the principles and procedures governing the processing of personal data with the aim of protecting individuals’ privacy, data security, and fundamental rights and freedoms.

Small Language Model (SLM): A natural language processing model that requires relatively fewer parameters and lower computational resources and is typically optimized for specific tasks or domains.

Training Data: The dataset used for artificial intelligence models to learn from.

2. Essential Principles
The principles listed here are part of Özyeğin University's AI Policy.

•  Artificial intelligence must be used ethically, safely, and responsibly, in compliance with international , , national , 4 and institutional regulations and principles.
• In all uses of AI, compliance with Law No. 6698 on the Protection of Personal Data (KVKK) and other applicable national and international data protection regulations must be ensured.
• Platforms and tools that may violate the privacy of personal data should be avoided.
• Confidential documents or information must not be uploaded under any circumstances to publicly accessible AI systems, including large language models (LLMs), small language models (SLMs), or AI-powered applications (e.g., ChatGPT, Copilot, Claude, DeepSeek), unless explicitly approved by the University and compliant with applicable information security requirements. The Information Security Office should be consulted before sharing any information or data that may be sensitive.
• The use of AI tools should be appropriately disclosed, including the tool(s) used, their function(s), the stage(s) at which they were used, and the extent of their contribution.
• Course instructors should clearly and comprehensively specify the guidelines they adopt for the use of AI in their course syllabi.
• The use of AI in teaching, research, administration, and community engagement must comply with applicable legislation5, 6, ethical guidelines, and University policies
• Users remain fully responsible for their use of AI tools, the outputs generated, and any decisions or actions based on those outputs.
• When using AI tools, users must adhere to principles of information security, data privacy, data minimization, ethics, accuracy, impartiality, and academic integrity.
• Artificial intelligence tools must not be used for cyberbullying, peer bullying, or the creation of deepfakes under any circumstances.
• AI tools may generate inaccurate, incomplete, misleading, biased, or discriminatory content. Accordingly, users are responsible for verifying the accuracy, completeness, reliability, and appropriateness of AI-generated outputs before using or sharing them.
• The use of AI tools in educational, academic, and administrative processes at the University must strictly adhere to any limitations imposed by the relevant departments or authorized boards.
• Records relating to the use of enterprise AI tools may be technically processed for system security, usage optimization, and cost analysis purposes. These records are stored in accordance with personal data security principles and are accessible only by authorized personnel.
• Any incidents arising from the use of AI tools, including, but not limited to, the accidental uploading of corporate data, excessive information disclosure, insufficient data masking, cybersecurity vulnerabilities, or data theft that may create administrative, legal, or technical risks, must be reported to the Information Security Office (bilgiguvenligiofisi@ozyegin.edu.tr ) within 24 hours.

3.1. Responsible Use of AI
The principles governing the circumstances under which members of the OzU community may or may not use artificial intelligence are set out under separate sections below.

Students
Students may use AI in accordance with this document, applicable national and international legislation, and Özyeğin University’s AI Policy, provided that:

• the course instructor has granted written permission; and/or
• the course syllabus explicitly permits the use of AI.

Such use must be lawful and limited to purposes such as those listed below.

• Finding resources, summarizing, and explaining concepts,
• Developing ideas, drafting content, and generating title suggestions,
• Preparing study plans and receiving reminder alerts,
• Translating, finding appropriate terms and expressions, and understanding texts,
• Preparing learning materials such as videos, podcasts, and presentations.
• Improving grammar, spelling, and fluency, except in ScOLa courses, assignments, and examinations.

The use of artificial intelligence tools in the following cases is not permitted without the prior written permission of the course instructor and/or unless explicitly specified in the course syllabus:

• The use of AI in examinations,
• The use of AI in the preparation of assignments, projects, and practical tasks, whether they affect the course grade (evaluation and assessment is at the discretion of the relevant instructor),
• Submitting content such as ideas, abstracts, and translations generated by AI without verifying their accuracy and originality,
• Failure to disclose the use of AI in the appropriate section of academic work.

Faculty (Teaching Staff)
Faculty members may use AI in accordance with this document, applicable national and international legislation, and Özyeğin University’s Artificial Intelligence Policy, provided that such use is lawful and limited to purposes such as those listed below.

• Designing or developing course syllabi,
• Developing suggestions for learning activities,
• Preparing learning materials, such as videos, worksheets, and podcasts,
• Getting help or suggestions regarding questions and assignments for evaluation and assessment,
• Providing guidance on the safe and ethical use of AI in the preparation of student assignments, projects, and practical tasks,
• Receiving support with translation, grammar, and spelling.

Faculty members should avoid using AI in the cases exemplified below:

• The uncontrolled delegation of critical tasks to AI, including academic advising or personalized student feedback,
• The preparation of course or evaluation-assessment materials (e.g. visual, auditory, written, etc.) entirely by AI without any intellectual input,
• The grading of examinations and assignments, as well as the provision of feedback, conducted entirely by AI,
• Presenting AI-generated content as original academic work,
• Failure to disclose the use of AI in the appropriate section of academic and other works.

Academic and Administrative Staff
All employees may use artificial intelligence in accordance with the principles set out in this document, applicable national and international legislation, and Özyeğin University’s AI Policy, provided that such use is lawful and limited to the purposes exemplified below:

• Utilizing AI for routine administrative tasks such as general document management and appointment scheduling, by ensuring the accuracy and confidentiality of outputs,
• Using AI for the analysis of reports containing objective and factual data, such as statistics,
• Providing rapid responses in situations that do not require discretionary or operational decision-making,
• Preparation of audio, visual, and written content for university promotion (with appropriate attribution),
• Developing presentations, graphics, audio content, videos, and motion graphics (with appropriate attribution),
• Supporting accessibility features such as translation, subtitling, and voiceover.

It is not recommended for employees to use AI in the cases exemplified below:

• Processing the University’s institutional or confidential information and documents without the approval of the relevant administrator,
• Unlawful processing of information and documents containing personal data,
• Using AI in matters or cases requiring subjective evaluation,
• Failure to disclose the use of AI in the appropriate section of administrative work.

Given the rapid evolution of AI technologies, applicable legislation, institutional policies, and best practices may change over time. Members of the Özyeğin University community are therefore encouraged to consult the most current versions of this document, the University's Artificial Intelligence Policy, and other relevant regulations before using AI tools.

1 OECD AI Principles: https://www.oecd.org/en/topics/sub-issues/ai-principles.html
1 EU General Data Protection Regulation (GDPR): https://gdpr.eu
3 Personal Data Protection Law (KVKK): https://www.mevzuat.gov.tr/mevzuat?MevzuatNo=6698&MevzuatTur=1&MevzuatTertip=5
4 Cyber Security Law: https://mevzuat.gov.tr/mevzuat?MevzuatNo=7545&MevzuatTur=1&MevzuatTertip=5
5 Guideline on the Responsible and Reliable Use of Generative Artificial Intelligence (GenAI) in Research Funding Processes: https://tubitak.gov.tr/en/institutional/about-us/generative-artificial-intelligence-guideline
6 Ethics Guide of Generative Artificial Intelligence Use in the Scientific Research and Publication Process of Higher Education Institutions: https://proje.yok.gov.tr/documentFiles/17539645794.Y%C3%BCksek%C3%B6%C4%9Fretimde%20%C3%BCretken%20yapay%20zeka%20kullan%C4%B1m%C4%B1-en.pdf