Generative AI and Pedagogy at SMU
- GPT (OpenAI). A versatile large language model capable of engaging in human-like conversations, answering questions, and assisting with various writing tasks such as tutoring, drafting, and brainstorming. It’s widely used to support teaching, research, and writing activities.
- Claude (Anthropic). An AI assistant known for its strong reasoning capabilities and ability to handle complex tasks. Claude excels at analysis, writing, and coding, making it useful for academic research and content creation. It supports educators by helping with content generation, curriculum development, and brainstorming ideas.
- Bing Chat / Copilot (Microsoft). Integrated within Microsoft Office products, this AI assistant helps faculty streamline tasks like writing, research, grading, and organizing class materials. It also enhances search capabilities for more effective information retrieval.
- Gemini (Google). Google's multimodal AI model capable of understanding and generating text, images, and code. It can be particularly useful for interdisciplinary research and projects involving multiple types of data. Google’s AI tool is designed to facilitate seamless integration with search and productivity apps, enabling personalized assistance in research, content creation, and problem-solving.
- MetaAI (Meta/Facebook). MetaAI focuses on large-scale language models that power social interactions and educational tools, helping educators engage students with personalized experiences and interactive learning resources.
- DALL·E (OpenAI). An AI model that generates images from text prompts, useful for creating custom visuals for lectures, presentations, and learning materials. It offers flexibility in content creation and artistic design for academic contexts.
- Stable Diffusion (Stability AI). A text-to-image generation tool that supports creating high-quality, realistic images for educational and artistic purposes. Faculty can use it to generate illustrative content for courses or research presentations.
- MidJourney. Known for producing high-quality, stylized images, MidJourney is used for academic projects that require unique and visually appealing graphics. It’s ideal for creative disciplines or enriching lecture visuals.
- Firefly (Adobe). Integrated into Adobe’s Creative Cloud, Firefly uses AI to generate images and enhance creative workflows. Educators can use it to quickly produce graphics, layouts, and design elements for teaching materials.
- Canva’s Magic Studio. A user-friendly AI-powered design tool that assists educators in creating professional-quality presentations, infographics, and visuals with ease. It simplifies graphic design for non-designers, making it accessible for all teaching disciplines.
- Semantic Scholar. A free, AI-powered research tool that helps professors find relevant academic papers quickly by filtering results based on significance and citations. It’s excellent for staying current with scholarly literature and it can significantly streamline literature reviews and research discovery.
- Elicit. An AI-driven tool that assists in academic research by summarizing and synthesizing scientific papers, providing insights, and suggesting relevant studies. It’s particularly useful for literature reviews and identifying key research gaps.
- Connected Papers. A research mapping tool that visualizes connections between academic papers. It aids professors in discovering related work in their field, supporting a more comprehensive understanding of existing research.
- Research Rabbit. Similar to Connected Papers, Research Rabbit provides visual representations of scholarly work, but focuses more on personalized recommendations and curated paper collections, helping researchers track trends and expand their literature searches.
- Scite.ai. This tool tracks citation contexts, showing how papers have been used or critiqued. It’s especially valuable for professors who want to see how their work is being cited in academic discourse or find papers with high scholarly impact.
- Zotero. While not primarily an AI tool, Zotero is a reference management software that now incorporates AI features for improved citation management and research organization. It can enhance the efficiency of academic writing and bibliography creation.
SMU Generative AI Policies
It is at the discretion of the faculty member whether to use Generative AI in their classroom. SMU requires that faculty communicate to students about the use of AI via their syllabus, and SMU has developed a statement builder for this purpose.
Faculty who use AI in the classroom should provide clear guidelines around AI use in the class for graded and ungraded assignments, self-study outside of class, and support through other SMU avenues, as well as a disclosure statement if AI is used for other pedagogical aspects such as assignment building or grading.
Faculty are asked to become familiar with SMU’s guidance around privacy and data security. It is important to remember that having access to data does not imply permission to scrape the data or use it to train an AI model.
Faculty are encouraged to consult SMU's AI-specific guides on topics such as "Managing Questions about Academic Integrity & Generative AI”; "Guide for Research & Data Analysis"; "Practical AI for Teaching"; "Streamline Your Workday – Using AI to Save Time"; "Practical AI for Creatives"; and "Leveling Up – From AI Beginner to Expert."
SMU is in the process of working with various departments on campus to gauge how AI can be best integrated in their curriculum.
Online Resources for Using ChatGPT in Your Class and Beyond
The following links will connect you with videos and articles detailing how you can use ChatGPT. Need a beginner tutorial? You'll find a great one here. Ready for something more advanced or developed specifically with university faculty in mind? We've got you covered.
| What You'll Learn | Link |
|
"Quick Start Guide to AI & Writing" |
| What You'll Learn | Link |
|
"ChatGPT 101" |
| What You'll Learn | Link |
|
"Advanced GPT Prompt Tutorial" |
| What You'll Learn | Link |
|
"Elevate Your Teaching with ChatGPT" |
|
"Create Engaging Lessons with ChatGPT" |
| What You'll Learn | Link |
|
"Professor Initiates Integration of ChatGPT in Classroom" |
|
"Why ChatGPT Is Both Exciting and Unsettling for Students, Faculty" |
| What You'll Learn | Link |
|
LINK TO RESOURCE |
|
LINK TO RESOURCE |
|
LINK TO RESOURCE |
|
LINK TO RESOURCE |
|
LINK TO RESOURCE |
|
LINK TO RESOURCE |
|
LINK TO RESOURCE |
|
LINK TO RESOURCE |
Q&A with Dr. Jennifer Culver - Senior Academic Technology Services Director - Simmons School of Education and Human Development
Faculty may have several worries about students using Chat GPT, including:
- Academic Dishonesty: One of the biggest concerns is that students may use Chat GPT to cheat on assignments, tests, or exams by relying on it to generate answers for them. This can undermine the academic integrity of the institution and negatively impact the learning outcomes for both the students and the faculty.
- Lack of Critical Thinking: Another worry is that students may become too reliant on Chat GPT and may not develop critical thinking skills or the ability to solve problems independently. This can limit their creativity, problem-solving skills, and ability to think for themselves, which are essential skills for success in many fields.
- Limited Understanding: While Chat GPT can provide helpful information, it may not be able to provide the depth and breadth of knowledge that students need to fully understand complex topics or concepts. Relying too heavily on Chat GPT can lead to a limited understanding of the subject matter.
- Bias: Another concern is that Chat GPT may reflect biases and stereotypes present in the data it was trained on, which could perpetuate inequalities or reinforce problematic beliefs or assumptions.
- Privacy: Finally, faculty may worry about the privacy implications of using Chat GPT. Depending on how it is used and implemented, it may collect and store data about students and their interactions, which could be a potential privacy risk.
- Focus on application and analysis: Design assignments that require students to apply what they have learned to real-world scenarios or analyze complex problems. These types of assignments cannot be easily completed by relying solely on Chat GPT, as they require critical thinking, analysis, and problem-solving skills.
- Personalized Assignments: Assignments that require students to draw on personal experience or conduct original research can also discourage reliance on Chat GPT. For example, students could be asked to write a reflection paper or a case study based on their own experiences or observations. This would require them to use their own insights and knowledge rather than relying solely on a machine-generated response.
- Use Rubrics: Use rubrics to clearly outline the expectations for assignments and include specific criteria related to critical thinking, analysis, and original thought. This will encourage students to think independently and provide original work rather than relying on Chat GPT generated responses.
- Use Plagiarism Detection Tools: Use plagiarism detection tools to check the authenticity of student work. This can help identify assignments that rely too heavily on Chat GPT generated responses and deter students from submitting entirely machine-generated work.
- Check for Consistency: When grading assignments, check for consistency in the quality and tone of writing. If there is a sudden shift in writing quality or tone from one assignment to the next, this may indicate that a student has relied heavily on Chat GPT-generated responses.
Overall, the key is to design assignments that require students to use critical thinking and analysis skills, and to provide clear expectations and guidelines for the assignment to ensure that students are producing their own work.
- Automated grading tools: Chat GPT can be used to develop automated grading tools that can grade assignments, tests, or exams. These tools can help reduce the workload for faculty, save time, and provide consistent grading across assignments.
- Research assistance: Chat GPT can be used to develop research assistance tools that can help faculty locate and analyze relevant research articles or data. This can save time and improve the quality of research.
- Curriculum development: Chat GPT can also be used to develop tools that can assist faculty in developing course curricula. For example, a tool could suggest relevant readings, assignments, or learning activities based on learning objectives or student needs.
- Communication tools: Chat GPT can be used to develop communication tools that can help faculty communicate with students more effectively. For example, a tool could send automated reminders about assignments or provide updates on course material.
- Assessment tools: Chat GPT can also be used to develop assessment tools that can help faculty evaluate their teaching effectiveness. For example, a tool could provide feedback on areas where faculty can improve or suggest alternative teaching methods.
Overall, Chat GPT has the potential to assist faculty in various aspects of their work, from grading to curriculum development, and help improve the overall quality of teaching and learning.
Yes, faculty can use Chat GPT to write emails or other letters. Chat GPT is a language model that is designed to generate text based on the input provided to it. This means that it can be used to generate text for a wide range of purposes, including writing emails, letters, and other types of documents.
However, it's important to note that while Chat GPT can generate text, it may not always produce high-quality or accurate content, particularly when it comes to communication that requires a personal touch. Faculty should use Chat GPT-generated content as a starting point and make necessary modifications to ensure that the content is appropriate for the intended audience.
Additionally, when using Chat GPT to generate content, faculty should be mindful of plagiarism concerns. If the generated text includes content from other sources, it's important to ensure that the content is properly cited and attributed to the original source. This can help avoid potential plagiarism issues and ensure that the generated content is original and meets academic integrity standards.
Overall, while Chat GPT can be a useful tool for generating content, it's important for faculty to exercise caution and use their judgment when using the tool for communication purposes.
The one thing I would want faculty to know about ChatGPT is that it is an incredibly powerful tool that can enhance teaching and learning in various ways. ChatGPT is a state-of-the-art language model that is capable of generating high-quality text based on the input provided to it. This means that it can be used to develop a wide range of educational tools, such as chatbots, writing assistance tools, language translation tools, personalized learning tools, and assessment tools.
However, while ChatGPT can be a valuable resource, it's important to use it responsibly and with care. Faculty should be aware of the potential limitations of the tool and exercise their judgment when using it. They should also be mindful of the potential ethical concerns surrounding the use of AI in education and ensure that they are using the tool in a manner that is consistent with academic integrity standards.
In summary, ChatGPT is an incredibly powerful tool that has the potential to revolutionize the way we teach and learn. However, it's important to use the tool responsibly and with care, and to be mindful of the potential ethical concerns surrounding its use.
The Center for Teaching Excellence is Here for You!
If you are interested in a one-on-one consultation on integrating AI into your teaching or exploring how Generative AI can make university teaching more fulfilling, request a consultation today.
The Center for Teaching Excellence (CTE) has organized numerous workshops on AI and pedagogy. Recordings of some of these workshops are available on the CTE YouTube page.
GPT Tools: Protecting Privacy & Data
The expanding use, availability, and popularity of Generative Artificial Intelligence (AI) engines such as ChatGPT, Dall-E, Bard etc. generates significant interest and concerns for higher education. Along with these opportunities and concerns,there are some critical threats and important security implications that all SMU communitymembers should take into consideration when using GPT tools.
While SMU will continue to review and consider more specific updated policies as this emerging field grows, all SMU community members should be aware of following existing best practices for data and privacy protections when using GPT or any technology systems. Publicly available generative AI systems are not secure or private!
As general guidance, prompts/queries submitted must comply with the following best practices:
- Never submit any account information (usernames or passwords)
- Never submit any Personally Identifiable Information (PII) or other sensitive or confidential material
- Never submit queries or prompts that would lead to issues if they were to be made public
- Never submit anything that includes your own or SMU’s Intellectual Property (IP)
- Always follow the relevant existing SMU policies on Data Protection, Acceptable Use, FERPA, Academic Integrity, Attribution and Ethics
- Never use public Generative AI tools to train AI models for people/human subjects unless those individuals have explicitly agreed to such usage and that the use of the data complies with SMU policies
This guidance is intended to include not just ChatGPT, but to all Large Language Models (LLMs) and AI chatbots including ChatGPT, GPT-n/x, Gemini, LLaMa, BLOOM etc.
All SMU community members are advised to take into consideration that all queries/prompts made to an public AI engine are collected and retained by the organization providing the AI service. Intellectual Property (IP) can be lost by inappropriate sharing.
All SMU community members are advised to take into consideration that AI models are designed to return responses that appear convincing, and all output should be viewed from this position and not as authoritative. LLMs have,in many cases,been trained on a large corpus of material, and as such there is potential that the provenance of material and data reproduced by the models may be obscure, carrying the potential risk of copyright or other IP infringement. All output must be independently validated.
These guidelines apply to public cloud-hosted Generative AI tools, Generative AI tools or AI models hosted by SMU researchers will provide their own security disclaimers and should be appropriately developed with SMU’s IRB’s policies.