Interview with A. Hasan Sapci

ChatGPT, a new generative AI platform, was made available to the public in November 2022. It has astonished the world, with many academicians shocked by its abilities. ChatGPT can write essays, research papers, emails, letters of recommendation, study guides, discussion prompts, rubrics, presentation scripts, and test questions with simple commands. Students can now generate cohesive and consistent text with just a few clicks, skipping the traditional essay-writing steps. They can also copy and paste all the content without being detected by plagiarism detection tools.

The introduction of calculators, Wikipedia, or WolframAlpha has sparked similar discussions in the past. However, ChatGPT and AI-based text generators are more than just “learning tools.” They use reinforcement learning to improve performance and train the chatbot to generate realistic responses.

The Provost’s Office at Adelphi University established a multidisciplinary AI Task Force to provide guidance on generative AI tools, conduct in-depth evaluations, and provide clear and practical guidance to the university community on their responsible use. The Task Force actively seeks and incorporates input from faculty to ensure a diverse range of perspectives and expertise in decision-making. We also work to identify ethical implications arising from the use of generative AI technologies. The Task Force will work closely with the College Committees, Academic Standards Committee, Committee on Academic Integrity, Committee on Teaching and Advisement, SCAIT, and University Writing Coordinator.

The Task Force prepared its first report for the Board of Trustees in February and also presented it to the Faculty Senate in the spring. The report by The Task Force on the Implications of Artificial Intelligence-Based Content Generators explores the critical thinking capabilities of generative artificial intelligence like ChatGPT, discusses the potential impact of generative AI on higher education, analyzes technical and legal considerations for developing a super-plagiarism tool, and provides best practices and recommendations for the Adelphi community, including faculty members and the IT department, on effectively utilizing AI and addressing challenges related to AI-generated content.

AI Multidisciplinary Task Force

• Committee on Teachin and Advisement
• University Writing Coordinator
• Academic Standards Committee
• College Committees
• SCAIT
• Committee on Acadmic Intergrity

At Adelphi University, we never considered banning the use of new generative AI tools, but we recognize that these tools can be used to plagiarize. We are working with cross-functional teams to develop new policies and guidelines for AI tools. The Faculty Senate Committee on Teaching and Advisement and the Faculty Center for Professional Excellence (FCPE) recently developed an updated syllabus template that includes updated language regarding AI-generated texts and shared it with the faculty members.

“Copying and pasting from any source into your assignments or exams without quotation marks, citations, and references constitutes plagiarism. Students are expected to produce and submit original work and to cite all sources appropriately. Unauthorized collaboration on any work, or the presentation of someone else’s work as your own, is plagiarism. Content generated by an Artificial Intelligence third-party service or site (AI-generated content, e.g. ChatGPT) without attribution or authorization is also a form of plagiarism. Unless explicitly stated, artificial intelligence-based technologies, such as ChatGPT or word mixing software, cannot be used to generate responses (partial or otherwise) for student assignments or exams.”

Traditional plagiarism detection tools such as Turnitin and SafeAssign compare submitted documents to a database of web pages, student papers, and peer-reviewed academic journals to identify plagiarism. These tools are designed to be highly accurate and can detect even small similarities between a submitted document and other sources. However, they cannot detect AI-generated text.

Turnitin recently added a feature to detect text generated by AI. However, the accuracy of the results can vary depending on the quality of the input data and the specific algorithm being used.

We are developing practical guidelines and policies to help students and faculty understand how to use AI tools ethically and responsibly. We will share these guidelines with the Adelphi community once they are finalized.

We believe that AI tools can be a valuable resource for learning, but they must be used to uphold academic integrity. We will continue to work  with our students and faculty to ensure this happens.

In 2023, we have seen a rapid acceleration in the development and adoption of AI technologies. Several companies have launched new products based on large language models (LLMs). By incorporating AI tools into the curriculum, students can be exposed to cutting-edge technology and use them as a research tool to enhance their learning experiences. This is an excellent strategy to equip them with valuable skills for the future.

However, it is also possible to use generative AI tools as plagiarism tools. The text generated by AI tools is very similar to human-written text, so students could submit an assignment without putting any effort and obtaining any new skills. Therefore, it is essential to approach generative AI cautiously, especially concerning academic integrity.

Adelphi University has already implemented a new syllabus template for AI-generated texts and shared it with all faculty members. We plan to integrate AI tools into our curriculum, train our faculty on identifying AI-generated text and educate our students on academic integrity.

Students are required to produce “original content,” and the use of AI algorithms to generate academic content (AI plagiarism) is not allowed. Therefore, we encourage the use of Turnitin’s new AI feature as a screening tool. However, we do not advise it as the only method to detect plagiarism. If Turnitin detects AI-generated text, other evaluation and assessment criteria should also be used to ensure that AI-generated text is thoroughly evaluated.

AI Task Force will provide comprehensive resources to faculty members on AI detection and best practices. This will include an online tool demonstrating the best techniques for AI detection and a new AI rubric that can be added to existing rubrics. AI tools should be used responsibly and ethically in the classroom. We believe that AI can be a valuable tool for achieving this goal. Instead of solely relying on AI reports for disciplinary actions, instructors can use AI detectors as a starting point for further investigation.

Educating students about the limitations and possibilities of AI detectors will help avoid unnecessary penalization, as these tools may produce false positive and false negative results due to their reliance on pre-defined rules and patterns.

By combining AI technology with other assessment methods, developing comprehensive AI-friendly rubrics, and focusing on education rather than punitive measures, institutions can harness the full potential of AI while maintaining academic rigor and integrity.

The Adelphi University AI Task Force was established to foster the integration of AI across various disciplines and promote collaboration in AI-related research and education. We aim to build a collaborative environment that encourages faculty members from different departments to work together on AI projects and exchange knowledge. We are committed to developing comprehensive guidelines for the ethical use of generative AI in teaching. These guidelines will align with the field’s best practices and uphold our university’s values.

The Task Force will also provide opportunities for faculty members to enhance their understanding of AI concepts and applications through workshops, seminars, and access to relevant resources.

The AI Task Force developed an AI rubric template for faculty members. Each undergraduate or graduate program might build their own version using this template. The AI rubric comprises eight criteria: coherence, repetitive language, lack of personality, use of uncommon language, repetitive and formulaic writing, overuse of technical language, inconsistent voice or tone, and verification of sources and citations.

We included detailed descriptions for each criterion and a rating scale ranging from 1 to 5. The AI rubric promotes objective evaluation and constructive feedback. Additionally, it encourages coherent and engaging content. The underlying principles are supported by linguistics, writing studies, and AI development research. Studies have shown that cohesive organization, concise writing, consistent voice, and accurate citation of sources contribute to better readability and credibility of written content.

With faculty members’ feedback, we will continue to improve this rubric and add additional recommendations for our faculty. At the same time, we will follow other academic institutions’ best practices and integrate them into our curriculum.