AI Tools to Promote Student Learning and Success: Examples
This section offers a variety of examples of how AI tools can offer valuable learning opportunities in both the humanities and in STEM, including in-class activities and sample assignments that can be used across the disciplines. It also considers accessibility and how AI tools like speech recognition or closed captioning can help diverse learners. These developments are not fully inclusive yet (for example, there are some free platforms, but they’re limited) but that will likely change as the technology continues to advance.
Some general uses of AI include having students evaluate AI-generated content in small group discussions, or use AI to help them organize their own ideas. AI can be used for brainstorming and for proofreading. AI can also be used to create practice exercises, quizzes, or study guides, which give students the opportunity to reinforce their own learning. Consider a major like computer science, for example. Xin Ye (Assistant Professor, Rockland Community College) plans to integrate AI into her courses in the fall 2023 semester. Students can use AI to receive feedback or corrections while they are coding.
In-Class Activities
Here are four sample in-class activities created by Racheal Fest (SUNY Oneonta; The Faculty Center for Teaching, Learning, and Scholarship Specialist in Pedagogy) that would introduce students to AI literacy in a variety of disciplines:
Example 1: Prompt Engineering
This activity invites students to experiment and evaluate a range of prompting strategies with a particular output in mind. Students begin by prompting ChatGPT or Bing to generate a scholarly or popular essay on an assigned topic, likely one derived from shared course content (say, an example of panopticism inspired by Michel Foucault’s well-known chapter).
Students should ask the AI tool to modify, correct, revise, or take a different approach to its output (for instance, give me a different example, or simplify your explanation, or break this down into steps. Online resources, including Wikipedia, list prompt intervention strategies). After students elicit several rounds of changes from the AI tool, they print a record of their engagement.
Finally, students can reflect on their process by assessing the efficacy of their own prompts and the quality of AI outputs. Did they manage to produce desired results? How? What strategies did they employ? How did they struggle? What does AI do well? What does AI do poorly?
Students could begin this work collectively, guided by an instructor who might elicit prompts and enter them into ChatGPT, Bing, etc., on a projector. After modeling this process and discussing prompting strategies, students could go on to complete individual assignments.
Example 2: Evaluating Form and Content of AI Outputs
This activity invites students to think critically about the strengths and challenges of content generated by both AI and other humans. First, students respond to a writing prompt relevant to course content without engaging generative AI.
Once they have completed the exercise, students enter the same writing prompt into ChatGPT, Bing, or another generative AI tool. Students then evaluate the texts, comparing and contrasting their formal features, their approaches to content, and more. What are the strengths and areas of growth for each? What differences and similarities do students notice? How does AI write? How do humans write? How could each improve their compositions and outputs?
Utah Tech University’s Center for Teaching and Learning provides additional examples of rethinking learning outcomes with AI. Their drop-down on Integrating AI in Teaching offers several good suggestions. For example, instead of asking nursing students to summarize research on evidence-based practice, they might compare an AI-generated summary and a summary written by a human to see what the differences are (and therefore, how that could change the outcomes of that scenario).
Example 3: Using AI for Brainstorming and Idea Development
AI can be used as a brainstorming tool to help undergraduates narrow down their ideas for research or to come up with initial ideas.. In a recent Alchemy webinar, “Harnessing the Power of AI: Transforming Assignments and Assessments in Higher Education”, Dr. Danny Liu (University of Sydney) suggested using ChatGPT this way: have the students encourage the chatbot to ask them questions and generate text based on their responses. The following example comes from a direct exchange with ChatGPT from Stephanie Pritchard’s (SUNY Oswego) English composition class:
Prompt: I have to write a research paper about a community I belong to and how it uses literacy for my English composition college class. Can you ask me some questions to help me brainstorm for the assignment?
AI: Certainly! Brainstorming is a crucial step in the research process. Here are some questions to help you explore how your community uses literacy and generate ideas for your research paper:
- What is the specific community you belong to, and how would you define it?
- How does your community value literacy, and what role does it play in daily life?
- What are the various forms of literacy within your community? (e.g., reading, writing, digital literacy, cultural literacy)
- Are there particular literacy practices or traditions unique to your community?
- How does literacy impact communication within your community? Does it influence how members interact and express themselves?
- Are there any challenges or barriers related to literacy in your community? (e.g., access to education, language differences, technological limitations)
The AI created 13 additional questions that we did not include here. Students can request that the AI ask one question at a time so they can respond to each one or provide a list so they can consider their options. And with more practice with prompt engineering, students can refine their questions and get even better at asking the AI for help.
Example 4: Using AI for Proofreading
Proofreading often gets neglected by students because it is time-consuming, boring, or they feel that there is not enough time to do it before an assignment is due. In addition to built-in spell checkers in programs like Microsoft and Google, many students also use Grammarly to help them with sentence structure, tone, etc. Students can use ChatGPT for proofreading as well.
For example, pasting a text into the chatbot and simply asking “Can you proofread this for me?” will yield different results than pasting in text with more specific requests like, “Can you identify the analysis in this history paper?” or “Please describe the counterclaim in this argumentative essay.” Students can then compare what the generative AI tool noted and see if they are achieving what they mean to in their writing.
AI can help students in other ways as well: for example, if students are writing in the sciences, they may be required to use passive voice. Asking ChatGPT to specifically check a text for active voice would be a helpful activity. Students can also ask ChatGPT to check their citations or to evaluate their introductory or conclusion paragraphs. However, they should be advised and reminded that generative AI can and does make errors and suggest erroneous information. Students must use their judgment before accepting generative AI suggestions.
Many of the major citation styles have adopted policies around citation and generative chatbots:
Citing generative AI in MLA Style
Writing Assignments
Here are two sample writing assignments created by Stephanie Pritchard (SUNY Oswego) that can be incorporated into a variety of disciplines:
Example 1: Reflective Process Book
While it will be very challenging now (read: almost impossible) to create out of class assignments that are “AI-proof”, some assignments might be easier for students to complete without assistance from AI. One example would be a semester-long reflective process book, which is an assignment that can be used alongside research. The sample process book assignment that’s linked here has a few notable features: it is graded mostly based on completion (this assessment strategy might lesson some of the grade anxiety faced by many undergraduate students), it is meant to be completed in stages (students will also have some time to work on this in class), and it is meant to be reflective (so students can describe what they learned but, more importantly, how they learned it). The process book asks students to think critically about their own research, discussions in class, and the steps they took to complete their research assignment. This assignment can be a strong addition to classes that require critical thinking, writing, and discussion. The process book assignment linked here is from an English composition class.
Example 2: Annotated Bibliography
There are also ways to integrate AI into research-based assignments, like an annotated bibliography. This annotated bibliography example, from an English composition class, asks students to find and evaluate six different credible sources for their research project. For every source, students must provide bibliographic information, a short summary, an evaluation of the source, and a brief discussion of how the source is relevant to their topic. There is a note at the bottom of this assignment which outlines what parts of the annotated bibliography can receive help from AI: to create appropriate citations, to help look for sources, and to help students understand the content of the sources they’re considering. Since annotated bibliographies are popular assignments, there are various ways generative AI tools can be used to simplify the process for students as they begin to learn how to conduct research, especially in introductory classes.
What we’ve learned from AI so far is that it can write reflectively – and convincingly – about experiences that it can absorb from the data on which it was trained but cannot actually have first-hand. The process book is meant to help students critically think about how they learned. It’s worth mentioning, too, that both of these assignments are part of a larger, semester-long scaffolded writing project.
Some additional out-of-class examples include rethinking reading response papers and other essay assignments.
AI Tools for Research Assignments
As generative AI tools proliferate, more special-purpose tools are being released for use. One such tool for annotated bibliographies and research-related assignments is elicit.org. Asking a question such as “What is the current interpretation of the cause of mid-latitude glaciers on Mars?” returns a list of relevant papers with bibliographic information and short summaries, along with a summary of the top papers which provides a snapshot of the field of research. Clicking on the title of any of the papers opens a new window with the full text of the paper, summary information, factors reflecting the trustworthiness of the paper, critiques listed in other sources, and related citations. The research landscape is likely to change rapidly with the further development of specialized generative AI tools.
Lab Reports
A typical college-level lab report structure includes an abstract, introduction, procedure, analysis, results, discussion, and conclusion. Some high school students may have experience with lab reports, but college-level writing is often more complex.
Dr. Trevor Johnson-Steigelman (Associate Professor, Finger Lakes Community College) suggests an activity you can introduce early in the semester by providing your students with a model lab report (this example can also be used to demonstrate your expectations and serve as a guide for your students).
Students can use generative AI to write abstracts or conclusions for lab reports by inputting their lab report’s main body into the generator. As we mentioned earlier, it can be a helpful activity for students to compare their own abstracts or conclusions to what the AI produces. AI can summarize key findings, and AI-generated results may offer valuable insights by helping students see what they may have missed.
As with many of the assignments and activities suggested here, it is important to encourage your students to continue to think for themselves instead of copying results into a final report: to focus on the process of evaluation and quality of sources, rather than the outcome alone.
Accessibility
The Accessibility Resources offices on many college campuses have seen a huge increase in numbers over the last couple of years. In addition to offering accommodations like extended time on exams, preferred classroom seating, and food allergy or dietary restrictions, many offices also have technology for students as well. Some examples of assistive technology include smart pens, recording devices, and speech-to-text software.
AI tools have begun to make big impacts on digital accessibility (the practice of designing and developing digital content that people with disabilities can use). This means that people with visual, hearing, cognitive, or motor impairments can have access to these tools, which promotes inclusivity and more equal opportunities. However, concerns about student use of AI tools to cheat and the policies that result could restrict the ability of disabled students to use AI tools to support their learning (McMurtrie, 2023).
Here are some examples of what AI tools can do (Alston, 2023):
- Create captions for videos
- Use speech to text or text to speech
- Test accessibility of various websites
- Offer language translation or transcription
Some of the tools are free (like Google’s Speech-to-text API and also Microsoft’s Azure Speech to Text) but many are still quite limited unless users choose to upgrade to a paid version.
Considerations for Online Classes
Teaching online, whether synchronously or asynchronously, involves many of the same issues concerning student use of generative AI as are found when teaching in-person classes. Syllabus statements are important to clarify acceptable uses of generative AI, and faculty should anticipate questions such as:
- What constitutes plagiarism?
- What is the policy on the use of AI tools like ChatGPT , Grammarly, GitHub Copilot, DALL-E, Google translate, etc.?
- How do you cite ChatGPT?
Rethinking the online course learning objectives and the related online assignments and assessments in the age of ChatGPT is now an essential aspect of preparing to teach online.
To best support online learners’ success, assignments should be designed to make critical thinking and the process of learning visible to the online course instructor and online classmates. Scaffolded assignments with feedback are more effective than just asking for a paper, or essay, as a completed final product in one step. These considerations will be especially important for asynchronous online classes. Synchronous online classes have the option to include Lane 1 activities (as described above) to assure student attainment of course learning outcomes through real-time activities and interactions. Asynchronous online courses have no built-in option for such Lane 1 activities and must rely more heavily on scaffolded learning activities that focus on the learning process more than the learning product.
Here are some recommendations for effective online teaching practices:
- Design assignments that must reference weekly course readings, content, and online interactions/discussions.
- If applicable, incorporate or ask online learners to incorporate very recent news and current events into assignments (there will be a lag in between the time a news event occurs and when it is added to AI language models, although how long is likely to decrease as time goes on).
- Prompt learners to incorporate their own personal views, experiences, examples, or aspirations into their assignments.
- Design assignments that build toward a final product, one step at a time.
- Allow online learners options in how they make their thinking and learning visible to you. For example, design activities where learners can create a multimedia response to an assignment.
