Skip to content Skip to main navigation Report an accessibility issue
High Performance & Scientific Computing

Abstracts AI in Higher Education Symposium


Presenter and Panelist Abstracts for UTK AI in Higher Education Symposium on September 22, 2023

10amAI and Equity
Presenter: Pavlo “Pasha” Antonenko
Title: Computer Vision, Sharks, and Fossils: Integrating AI in Teacher Education
Abstract
This presentation will discuss an AI in Education project funded by the National Science Foundation. The purpose of this project is to help in-service and pre-service teachers integrate AI in the classroom using fossil shark teeth, Google’s Teachable Machine and Roboflow. Current and future teachers learn to conceptualize and implement ML models that distinguish fossil shark teeth by their morphology and primary functions to recognize ecological and evolutionary patterns. Our curriculum is scaffolded using the machine learning development process: 1) data collection and preparation; 2) selecting and training the model; 3) evaluating the models’ accuracy; 4) tuning model parameters to improve performance and reduce bias. Each stage of the development process constitutes a different module and is discussed during a week-long summer professional development (PD). Through these lessons, teachers are introduced to AI methods and tools, including two platforms used to build/train ML models: Google’s Teachable Machine and Roboflow. Preliminary mixed-method data analyses show teachers’ self-efficacy around teaching AI improved after engaging in the summer PD. Longitudinal data collection is underway and will inform future work related to improving Title I teacher and student self-efficacy related to using AI.
Presenter: Kenneth Fleischmann
Title: Building Equitable AI with Good Systems: Ethical AI at UT Austin
Abstract
How can we build human-AI partnerships that lead to a more equitable, just, and harmonious society? This talk will provide an overview of Good Systems, an eight-year research grand challenge at UT Austin launched in 2019. The talk will discuss the need for Good Systems and how this research grand challenge was founded. It will describe the partnerships built up through Good Systems, such as a Master Interlocal Agreement with the City of Austin and a Master Research Agreement with MITRE. It will include findings and impacts from completed projects within Good Systems, including projects focusing on equity in areas such as AI for accessibility, cultural competency, and social services. It will also detail the six core research projects that make up the main mission for Good Systems as well as education and policy efforts that have grown from Good Systems.
Presenter: Hope Chidziwisano
Title: Leveraging Human-Centered Approaches to Design Inclusive AI Applications
Abstract
As each day passes, the adoption of artificial intelligence applications (AI) is increasing globally. Over 90% of companies today embrace AI applications to improve service delivery [2]. Despite numerous advantages, AI applications come with potential risks for marginalized communities. This is because AI applications and algorithms are primarily developed by non-representative populations, and then exported to marginalized communities [1]. The transfer of technology from one region to another results in applications grounded in nonrepresentative data, exacerbating existing social and economic inequalities. For example, ChatGPT—an AI chatbot that has gained exponential adoption—is heavily trained on data from Western societies. Between 2015 and 2020, data from only one African country—Egypt—was used to evaluate ChatGPT’s performance [3]. This makes it difficult for AI models to produce results that match different contexts’ cultural and economic realities. Relatedly, decision-making AI algorithms in the financial sector exclude women from accessing loans because of their low education status in African countries, such as Uganda [1]. If we aim to develop equitable AI applications for global use, we should use existing infrastructure to collect data that reflects global realities. In doing so, we would encourage inclusive approaches from the initial stages to implementing AI-based solutions. These approaches are inscribed in new emerging sub-fields, including human-centered AI, responsible AI, and human-AI collaboration. By extending these sub-fields to marginalized communities, we would include underrepresented users and local technicians who are well conversant with existing socio-cultural and infrastructural challenges. More importantly, inclusive approaches— in the presence of technological infrastructure—would also result in collecting vast amounts of representative data that can be used to train well-balanced AI models.
11amPresenter: M’Hammed Abdous
Title: Multimodal Generative AI: Transforming Content Creation in Higher Education
Abstract
Multimodal generative AI – advanced technology that seamlessly blends text, audio, video, and graphics – has the potential to transform content creation in higher education. This presentation will explore the practical ways in which this AI can enhance learning materials and drive fundamental changes in teaching methods, learner engagement, and institutional management. Join us as we examine the impact of AI on content creation in higher education. Explore how multimodal generative AI can enhance learning materials, fundamentally change teaching approaches, increase learner interaction, and streamline administrative processes. By exploring the intersection of pedagogical innovation and digital evolution, this session will navigate the ways in which generative AI can disrupt and enhance content creation across the higher education landscape.
Presenter: Kostas Vogiatzis
Title: Teaching Artificial Intelligence for Chemical Applications
Abstract
Artificial intelligence (AI) is a general purpose, enabling technology that is considered nowadays as important as electricity, transforming every industrial and scientific sector. AI rapidly changes many aspects of chemical sciences, from drug discovery and material design, till the introduction of robotics in chemical laboratories. Recognizing this unique opportunity, I taught during Fall 2021 an undergraduate course (CHEM470: Special Topics in Chemistry: Artificial Intelligence and Chemistry) that brought together chemistry with computer science, statistics, and data science. This course covered the key aspects of AI and modern chemoinformatics including molecular representations and descriptors, virtual screening of databases for drug and material discovery, machine learning (ML) for computational chemistry, and chemical text mining. The target audience for this course was undergraduate students with major/minor in chemistry, biological sciences, but it was also open to other disciplines as well (e.g. physics, mathematics, electrical engineering and computer sciences, and chemical and biomolecular engineering). Overall, it attracted 36 students.
In my talk, I will present the scope, organization, assignments, and outcomes of this course, together with challenges and limitations. For example, during the preparation of this course, we identified the lack of a textbook that covers molecular representations (vectorizations of chemical structure and chemical information), a vital aspect for successful development of chemical AI/ML. This challenge motivated us to write during 2022-2023 our own textbook that was published in 2023 under the ACS in Focus series of the American Chemical Society (ACS).
Presenter: Deloris C. S. James
Title: AI and Workforce Development: A Case Study
of an AI Training Program for Health
Education and Behavior Majors at the
University of Florida
Abstract
Health education is a rapidly growing field, and the U.S. Bureau of Labor Statistics expects the overall employment of health education specialists and community health workers to grow by 12% in the next decade. The primary drivers of this growth are an aging population, immigration, high rates of mental health disorders, and the high morbidity and mortality rates of chronic diseases, especially among vulnerable and underserved populations. Other contributing factors include the increased emphasis on preventive healthcare, reduced healthcare costs, and effective educational responses to global challenges such as pandemics. This presentation discusses the Department of Health Education and Behavior’s best practices for cultivating a modern workforce of certified health education specialists (CHES) knowledgeable and skilled in utilizing AI tools for superior patient engagement and improved health outcomes. The presentation outlines the iterative process for developing an on-demand training program, AI-enhanced courses, linkage with industry partners, and experiential learning through internships and practicums. This initiative aligns with the University of Florida’s quality enhancement program to integrate AI across the curricula.
12:20pmPresenter: Jiaqi (Jackey) Gong
Title: From Numbers to Narratives: Teaching Creativity
in the Data Science Classroom
Abstract
This talk underscores the crucial interplay between analytical acumen and the
craft of storytelling in data science education. Beyond mere data analytics
comprehension, it’s the delicate art of weaving data into captivating narratives that
enhances the overarching influence of data science. This talk also unveils a
groundbreaking project delving into advanced teaching methodologies that harmoniously
blend in-depth analytics with imaginative story construction, emphasizing the significance
of transforming raw data into compelling and insightful stories.
Presenters: Jason Johnston & John Nash
Title: Reimagining Online Assignments With and
Because of GenAI
Abstract
Higher education is undergoing significant transformation, with Generative Artificial Intelligence (GenAI) playing a disruptive role. This presentation aims to explore the dual impact of AI on online assignments: how the advent of GenAI influences the way students (and instructors) approach assessment, and how educators can integrate GenAI to enhance learning outcomes.
Students using GenAI tools approach assignments with a different mindset, often leveraging these tools to optimize their efforts. Sometimes these efforts are academically dishonest, but often students are acting in a new “grey” area. The first half of this presentation will consider how this new relationship with GenAI necessitates reevaluating traditional assignment structures and student expectations. Assignments should be reimagined to promote genuine comprehension and critical thinking, rather than mere task completion, and explained using clear language about GenAI use.
The second half of the presentation will consider the potential of embedding GenAI within assignments. By integrating AI-driven tasks, we can not only familiarize students with GenAI but also stimulate critical reflection on its application. This approach guides students to not only be consumers of GenAI but also thinkers who are more ready to navigate a future workplace where GenAI is increasingly used.
By attending this presentation, university faculty and administrators will gain insights into the evolving dynamics of student engagement in the GenAI era. The discussion will provide actionable strategies for online course design, ensuring that educational methodologies align with these new advancements and that students are more prepared for the future.
1:00pmPresenters: Anne-Helene Miller and Doug Canfield
Title: Using Conversational AI in a VR Classroom
to Explore the Late Medieval Anglo-Norman World
Abstract
In this presentation, Dr. Miller (Riggsby Director of the Marco Institute) and  Dr. Canfield (Director of the Language Resource Center) will discuss their project entitled “Using Conversational AI in a VR Classroom to Explore the Late Medieval Anglo-Norman World.” This project involves creating a 14th-century Anglo-Norman castle in VR populated with period-specific avatars programmed with conversational AI. Students can engage with AI figures like a noblewoman, a ghostly crusader, a troubadour, a peasant, and a nun to learn about aspects of medieval society. Through natural dialogue in French or English, students inquire, explore personal narratives of the characters, and understand challenges and beliefs of the time. AI avatars can also assign quests such as feasts and truce negotiations. Activities encompass Guided Group Tours, Role-playing, and Reflection Activities. This VR and conversational AI combination offers a sensory-rich platform, grounding historical concepts in tangible experiences. The presentation focuses on using AI for student engagement, technological innovation, AI tools in education, and faculty training for AI in education. 
Presenter: Alex Bentley
Title: AI in Education: Socratic Search of a Knowledge Landscape
Abstract
The most discussed example of generative AI used in education are the Large Language Models (LLM)–ChatGPT being the most known of all of them. These LLM, and future generative AI tools using Human reinforcement learning (RLHF), will undoubtedly change how students learn. As researchers in the field of cultural evolution and social learning, however, we would argue that this may not transform education as radically as some would think.  The processes of using LLM to access specific knowledge still implies a back-and-forth “conversation” with the AI that requires skills and knowledge of the student. Since generative AI functions encompass a huge landscape of digital human knowledge, astute students will use iterative questioning, refining and re-questioning to optimize their use of LLM.  Students querying LLM will still need knowledge of the subject, just like a student engaging in age-old Socratic dialogue with a teacher. This interaction loop between LLM queries and the next knowledge step has similarities to the evolution of human social learning. The process reflects a search of a knowledge “landscape” aimed at climbing local “peaks” of expertise. In short, the knowledge and skills that applied thousands of years ago learning from an expert will still apply today when interacting with LLM, on whatever the subject may be.
Presenter: Dan Harder
Title: AI Tools for Education from OIT
Abstract
Dan will present on the AI Tools for Education available and in development from OIT for the UTK community.
2:00pmPresenter: Asim Ali
Title: The Online, Self-Paced Teaching with AI Resource
Asim Ali, Lindsay Doukopoulos, Shawndra Bowers and DeElla Wiley
Abstract
At Auburn University, the Biggio Center for the Enhancement of Teaching and Learning has developed an online, self-paced resource titled “Teaching with AI.” The course has eight modules and covers topics such as the basic concepts of generative AI, ethical considerations, designing courses and assignments, partnering with students on appropriate AI uses, and empowering faculty to have conversations about AI within their disciplines. The course has over 600 participants at Auburn University. Recently it was opened to other campuses and is now used by hundreds of faculty at nearly 35 campuses, including the Southeastern Conference institutions as the result of a unique partnership between Auburn University and the SEC Academic Relations office. To the best of my knowledge, this is a unique offering in the faculty development landscape.
The presentation will cover the process of developing the course, the elements of the course, and the digital badge and how it is earned in the course.
Presenter: Claire Mayo
Title: CAS AI Educational Campaign – RFP:
Exploring the Use and Integration of
Generative AI Tools in Classroom Activities
Abstract
The College of Arts & Sciences, in collaboration with the AI Tennessee Initiative and Teaching and Learning Innovation, has initiated a comprehensive Request for Proposals (RFP). This RFP is designed to explore the seamless integration of generative AI tools within higher education, with a specific emphasis on their application in teaching, learning, monitoring, and assessment. Preliminary findings stemming from nearly two dozen expressions of interest have illuminated common themes. These encompass the development of robust assessment and monitoring methodologies utilizing generative AI, the creation of inventive classroom activities empowered by generative AI technology, the enhancement of students’ digital literacy concerning AI tool usage, and the consideration of discipline-specific opportunities and challenges, among others. At the symposium, we will delve into the insights gleaned thus far from this initiative, elucidating our approach to its execution. The primary objective is to foster engaging dialogues among educators, researchers, and AI enthusiasts. We aim to facilitate a comprehensive exploration of generative AI tools’ potential in reshaping education while conscientiously addressing ethical implications and challenges.

To understand CAS faculty perceptions and readiness regarding AI adoption, we conducted a Qualtrics survey in late July and early August 2023. We offer a concise overview of our findings from the survey in this talk at the symposium. Participants who had interacted with ChatGPT shared their experiences, offering valuable feedback on its utility in education. Faculty comfort levels in using ChatGPT for teaching and research were also assessed, revealing their willingness to incorporate AI tools into their academic activities. The survey delved into ethical considerations, identifying faculty awareness of AI-related ethical concerns and their interest in additional resources for navigating these issues. Overall attitudes toward AI language models like ChatGPT were explored, providing insights into faculty perceptions of AI’s potential to enhance the learning experience. Practical applications of ChatGPT in education, concerns about its impact on faculty workload, and its role as a supplement or replacement for human interaction were also examined through open-ended questions. Faculty suggestions for responsible AI use included education, training, and clear guidelines to ensure ethical and effective AI integration. Finally, faculty highlighted challenges and limitations related to academic dishonesty, misinformation, and student skill development, offering a comprehensive view of their concerns. These survey results provide valuable insights into faculty perspectives on AI in higher education. They serve as a resource for institutions aiming to tailor support, training, and policies to responsibly integrate AI into the academic landscape.
Presenter: Mehmet Aydeniz
Title: How to Engage, Educate and Motivate University Faculty to use GENAI Tools in Teaching and Research?
Abstract
The introduction of ChatGPT by OpenAI in November of 2022 has created fear, confusion and excitement, all at the same time, among higher education faculty and administrators. After OpenAI introduced ChatGPT other companies came up with several other GENAI tools that can serve diverse operational, pedagogical and research goals in higher education. While this technological development in the AI space is happening at a rapid pace, faculty members’ adoption of such tools for pedagogical and research purposes is not consistent with the pace of these technological developments. While the source of this behavior is complex and multifaceted it needs immediate attention. If we want to take advantage of these tools to increase pedagogical effectiveness and research creativity in higher education, we need to bring faculty on board. However, faculty members have been slow in adopting the GENAI tools for teaching and research purposes. In this talk, we discuss the source of this paucity in adoption and present a faculty training and engagement framework that holds promise for effective adoption of GENAI tools for research and teaching in higher education. 

The flagship campus of the University of Tennessee System and partner in the Tennessee Transfer Pathway.