Introduction to Generative AI (GenAI)

Image: AI icon

Introduction to Generative AI (GenAI)

AI & Teaching

What is it?

What is it?

Generative AI (often shortened to “GenAI”) is a form of artificial intelligence through which computers generate content by pulling, synthesizing, and reorganizing data drawn from a large dataset. 

GenAI uses algorithmic techniques called neural networks to scan large amounts of data, recognize patterns within that data, and then generate a new or different iteration of information based on those patterns. The datasets used to train GenAI are drawn from massive amounts of human-generated text and are analyzed with algorithms known as Large Language Models (LLMs). 

GenAI chatbots mimic the linguistic and content patterns they detected in human communication data, so their outputs tend to “sound” human. Because these machines also draw upon a vast quantity of human communication data, the output they produce can have tremendous sophistication in both content and language. 

Generative AI chatbots are not inventing new ideas, nor can they assess the accuracy, contextual relevance, or appropriateness of the output they produce. Chatbots are not “thinking,” they are pulling and re-synthesizing textual patterns learned from LLMs.

Research

Research

Research on the effect of GenAI on teaching and learning is rapidly underway. Emergent research trends show that GenAI holds both opportunities and limitations for learning. 

In their study on the applicability of ChatGPT for science education, Cooper (2023) found that, while ChatGPT is able to produce strong and coherent responses to scientific questions, it is unable to provide clear evidence for its outputs. Darics and van Poppel (2023) argue that one of the limitations of GenAI is its tendency to conflate popularity with truth: “The world ChatGPT presents to us is based on argumentum ad populum – it considers to be true what is repeated the most.”

While much of the popular and scholarly attention to GenAI in higher education has centered on issues of academic integrity, Dobrin (2023) argues: “The prevalence of these discussions foregrounds two worrisome assumptions…a) that students will always cheat if given the means, and b) that a primary function of instructors and institutions of higher education is to police students and their use of such tools” (12). Furthermore, as Guzman (2023) argues, students express a wide range of dispositions toward GenAI and many students have not adopted GenAI tools as shortcuts nor supplements to their learning. 

AI-aware teaching and learning efforts must teach students how to use Generative AI with integrity. As Eaton argues: “We would never deny students access to the internet, dictionaries or spell check. So why would we deny them access to this new tool?...If we expect students to act with integrity, then we as educators have to act with integrity and model that behavior… Nobody wins in an academic integrity arms race” (as quoted in Wilhelm, 2023).

Teaching Strategies

While we do not yet have a body of scholarship on best practices with AI, below are strategies for being “AI-aware” in your teaching based on sound pedagogical approaches and tested frameworks for bringing technology into the classroom.

  • Don’t ignore AI. AI is here and students are already using it. Your students may already be using GenAI to supplement their learning and their assignments in your courses.

  • Have an explicit, transparent, and consistent AI policy. Talk with students about how, when, and why they can use GenAI tools in your course.

    • Explicit: Put your policy in your syllabus and refer to it often.

    • Transparent: Clarify your thinking and rationale behind the course policy. Consider how use of GenAI can further and/or hinder your learning outcomes, and help your students understand how GenAI can support and/or limit their learning in your course.

    • Consistent: Hold all students to the same policy regardless of situation. Avoid extreme policies that you may have a difficult time enforcing later.

  • Give students clear guidelines on how to think about human-computer interaction and the ethical use of texts/outputs produced from that interaction. Encourage students to be transparent about their use of GenAI and all other tools they use to supplement their learning and assignments. Require students to cite GenAI output, describe the extent to which they used GenAI for an assignment, and reflect on the impact of that interaction for their learning.

  • Reconsider your approach to assignments and assessment. Rather than trying to AI-proof your assignments, try to integrate assessment practices that have been proven to increase student’s intrinsic motivation and promote deeper learning. Some suggestions:

    • Utilize frequent, lower-stakes assessments that emphasize building upon and/or revising work completed over time.

    • Embrace transparency in assignments by designing assignments with a clear and explicit purpose. Talk with students about why they are engaging with the assignment and what they will learn from it. Clarify your expectations for what a successful assignment will look like and/or be able to do, so that students can self-assess their own work appropriately.

    • Design assignments that require students to incorporate formative feedback, and ask students to explain how they revised previous iterations based on feedback.

    • Ground assignments in recent events or current issues, especially those affecting our local region and/or students’ home communities. Not only will these be more intrinsically relevant and motivating for students, these events will not have a large amount of associated data in the training datasets used by GenAI chatbots.

  • Engage with GenAI tools collaboratively and critically, as a class. These are new and rapidly changing technologies. You will not be, nor do you need to be, an authority on GenAI. You can still model expert-like thinking by showing students how to engage with and recognize the limitations of GenerativeAI. Some suggestions:

    • Consider incorporating “AI literacy” into your course learning outcomes.

    • Signal to students that you will be learning alongside them as you explore GenAI, and encourage your students to engage in the class as a community of learners.

    • Co-construct your AI policy with your students. Ask students what they want to learn about GenAI, what they think is an appropriate or inappropriate use of GenAI, and whether they are already using GenAI in their learning and lives.

    • Engage students in recognizing the flaws of GenAI chatbots and the risks of using GenAI output, including the ethical and social considerations detailed above.

    • Have students edit and revise AI-generated content to perceive the difference in human vs. machine-generated communication.

  • Do not require students to use GenAI or feed their ideas/drafts into GenAI chatbots. There are serious ethical and social implications to consider when using GenAI and many students may experience these tools as inaccessible, harmful, or unethical. Until these concerns are addressed, GenAI should be a voluntary, supplemental option for students, rather than a requirement.

  • Explore both the affordances AND the limitations of incorporating AI into assignments and other learning activities. Remember that GenAI chatbots cannot think, nor can they check the accuracy of their own outputs. However, they can provide novel or interesting outputs with which students can critically engage. Some examples of using GenAI in learning activities might include:

    • Have students use GenAI to generate multiple ways of expressing an idea, code, formula, or concept, then compare and contrast the relative strengths and weaknesses of each output.

    • Use GenAI to develop counterarguments to disciplinary claims or arguments, and ask students to assess the evidence presented in these counterarguments.

    • Have students argue for and/or against AI-generated output using the key topics, terms, readings, formulas/equations, techniques, and/or materials covered in the course.

Sample Syllabus Statements
 

The following samples provide a variety of approaches to including a GenAI statement in your syllabus.

  • If GenAI is permitted with attribution: 

  • The use of generative AI writing tools (such as ChatGPT, GrammarlyGO, GPT-3, GPT-4, Elicit, BERT, or others) to support you as a writer (e.g., for brainstorming, finding search terms for research, translating, getting feedback for revising and editing) is allowed in this class. Such use must be properly acknowledged in references, bibliographies, or other formats as required by the instructor and must be discussed in reflective writing (e.g., assignment memos).

    Learning to use AI in productive, responsible, and ethical ways to support writing processes is an emerging literacy skill, one that we will begin to cultivate in this course through readings, guided activities, discussions, and assignments. We will review and practice as a class how to properly acknowledge use of AI writing tools and properly cite and reference AI generated content.

    As we learn about AI and how AI writing tools can support different aspects of the writing process, we will be guided by two core principles:

    Cognitive dimension: Working with AI should not reduce your ability to think clearly. We will practice using AI to facilitate—rather than hinder—learning.
    Ethical dimension: Students using AI should be transparent about their use and make sure it aligns with academic integrity. (UC Davis University Writing Program)
    - - - 

    In principle you may submit AI-generated code, or code that is based on or derived from AI-generated code, as long as this use is properly documented in the comments: you need to include the prompt and the significant parts of the response. AI tools may help you avoid syntax errors, but there is no guarantee that the generated code is correct. It is your responsibility to identify errors in program logic through comprehensive, documented testing. Moreover, generated code, even if syntactically correct, may have significant scope for improvement, in particular regarding separation of concerns and avoiding repetitions. The submission itself should meet our standards of attribution and validation. (University of Michigan)

    - - - 

    Generative AI (Artificial Intelligence that can produce contents) is now widely available to produce text, images, and other media. We encourage the use of such AI resources to inform yourself about the field, to understand the contributions that AI can make, and to help your learning. However, keep the following three principles in mind: (1) An AI cannot pass this course; (2) AI contributions must be attributed and true; (3) The use of AI resources must be open and documented. (Sentient Syllabus Project)

  • If GenAI is not permitted: 

  • ChatGPT and other similar technologies are advancing rapidly and there are many instances where they will be key tools in your schoolwork and career. For the purposes of this class, we are asking all students to pledge that they will not use these technologies. We believe this is key for this learning environment because we want you to learn how to critically engage with the material we’ll be discussing, including learning how to search for and identify relevant sources, synthesize these materials, and make recommendations without the aid of technology. Artificial Intelligence cannot do this learning for you. Students who are found to have used ChatGPT or the like to complete their assignments will receive a grade of zero for that assignment.  (University of Michigan)

    - - - 

    In this course, all assignments must be completed by the student. Artificial Intelligence (AI), including ChatGPT and other related tools used for creating of text, images, computer code, audio, or other media, are not permitted for use in any work in this class. Use of these generative AI tools will be considered a violation of Academic Integrity Policy, and students may be sanctioned for confirmed, non-allowable use in this course. (Arizona State University)

  • If GenAI is conditionally permitted for some activities, but not others:

  • During our class, we may use AI Writing tools such as ChatGPT.  You will be informed as to when, where, and how these tools are permitted to be used, along with guidance for attribution.  Any use outside of this permission constitutes a violation of [Academic Integrity] Policy. (Bryant University)

    - - - 

    There are situations and contexts within this course where you will be asked to use AI tools to explore how they can be used.  Outside of those circumstances, you are discouraged from using AI tools to generate content (text, video, audio, images) that will end up in any student work (assignments, activities, responses, etc) that is part of your evaluation in this course. Any student work submitted using AI tools should clearly indicate what work is the student’s work and what part is generated by the AI... (Colorado University)

Students say …

Students Say

  • One survey of 1,200 current undergraduate and graduate students found that about one-third of students had used GenAI for schoolwork in the past academic year (Kyaw 2023).

  • Another survey from BestColleges found that 40% of students surveyed had already used GenAI in some capacity, and about 22% of students had used GenAI to complete assigned coursework or exams (Welding 2023). 

  • More than half of the students indicated that they considered GenAI a form of cheating, and over half of students reported that neither their schools nor instructors had specified how to use GenAI ethically or responsibly.

Reflection

Reflection

  • What are your initial thoughts on the possibilities of integrating GenAI into your teaching practices?
  • Do you have an AI policy that you are practicing with your students? If yes, how is it going? If not, what do you think the policy should be?
  • Have you explicitly signaled to students when, how, and why they can/cannot use GenAI in your course? 
  • How might you model ethical and responsible use of GenAI for your students, so that they can learn to use this tool with critical thinking and integrity?
  • How might you update your assignments and/or assessment practices to increase student’s intrinsic motivation and engagement with their coursework?

Further Resources

 

Dobrin, S. (2023). Talking about Generative AI: A Guide for Educators. Broadview Press: Ontario, Canada. 

University of Michigan. Course policies & syllabi statements. U-M Generative AI Guidance. Retrieved from https://genai.umich.edu/guidance/faculty/course-policies. 

Welding, L. (2023). Half of college students say using AI on schoolwork is cheating or plagiarism. BestColleges. Retrieved from: https://www.bestcolleges.com/research/college-students-ai-tools-survey/ 

Wilhelm, I. (2023). ‘Nobody wins in an academic integrity arms race. Chronicle of Higher Education. Retrieved from  https://www.chronicle.com/article/nobody-wins-in-an-academic-integrity-arms-race.

Contributors

Erica Bender, Katie Healey