EDVICE EXCHANGE

Dana Dawson

In a previous blog post in this series, we suggested making generative AI a subject of critical analysis in your courses. Here, we will focus on the importance of teaching our students to critically engage with content generated by AI tools and with the implications of generative AI use for our information environment. This topic lies at the intersection of digital literacy, information literacy and the newly emerging field of AI literacy (Ng et al.; Wuyckens, Landry and Fastrez). While our students will need to develop the digital literacy required to solve problems in a technology-rich environment characterized by the regular use of AI tools, they will require the information literacy skills to navigate through a complex information ecosystem. Though generative AI tools are digital tools that generate information, we have a tendency to interact with AI tools as if they are social beings (Wang, Rau and Yuan, 1325-1326) and the manner in which they generate information requires special attention to issues of authorship, the impact of data-set bias and the potential automation of disinformation dissemination.

• As the efficacy and availability of generative AI tools advances, both we and our students will face a variety of information-related challenges. Generative AI can be used to automate the generation of online misinformation and propaganda, significantly increasing the amount of mis- and disinformation we are exposed to online. Flooding our information environment with disinformation not only increases exposure to bad information, but distracts from accurate information and increases skepticism in content generated by credible scholarly and journalistic sources. Even where users do not intend to propagate misinformation, Ferrara and others have pointed out that bias creeps into text and images produced by generative AI through the source material used for training data, the design of a model’s algorithm, data labeling processes, product design decisions and policy decisions (Ferrara, 2).These limitations can result in the creation of content that seems accurate but is entirely made up, a phenomenon known as AI hallucinations.

Our task as educators is to prepare our students to navigate an information environment characterized by the use of generative AI by inoculating against disinformation, helping them develop the skill and habit of verifying information, and building a conception of the components of a healthy information environment.

Tools for Inoculation

Inoculating ourselves and our students against mis- and dis-information functions much the same as inoculating ourselves against viruses through controlled exposure. By “pre-bunking” erroneous content students may themselves create using generative AI tools or may encounter online, we can help reduce the potential for them to be misled in later encounters.

• Ask students to use ChatGPT to outline one side of a contemporary debate and then to outline the other side of the debate. Have them experiment with prompting the tool to write in the voice of various public figures or to modify the message for different audiences. Analyze what the tool changes with each prompt. Look for similar messages in social and news media.
• Use the resources of the Algorithmic Justice League to explore how algorithms reproduce race- and gender-based biases.
• If you assign discussion board entries to your students, secretly select one student each week to use ChatGPT or another generative AI tool to write their response. Ask students to discuss who they believe used AI that week and why.
• Have students experiment with Typecast or other AI voice generators to create messages in the voice of public figures that are aligned or misaligned with that individual’s stance on contemporary issues.
• Have students investigate instances of the use of tools such as Adobe Express to create misleading images that circulated online (for example, fake viral images of explosions at the Pentagon and the White House). The News Literacy Project keeps a list here. Analyze who circulated the images and why. How were they discovered to be fake? Ask students to experiment with the image generating and editing tools used in the instances they discover, or with free alternatives.

Tools for Verifying Information

Zeynep Tufekci argues that the proliferation of generative AI tools will create a demand for advanced skills including “the ability to discern truth from the glut of plausible-sounding but profoundly incorrect answers.” Help your students hone their analytical skills, understand the emotional aspects of information consumption and develop a habit of questioning and verifying.

• Increase students’ self-awareness of their own information consumption habits and methods for verifying information they are exposed to. Ask students to keep a journal for a week of their social media consumption and what they shared, liked, up- or down-voted, reposted, etc. on social media for a week. What kind of content do they tend to engage with? What feelings motivated them to share or interact with content and how did they feel afterward? If shared content included information or took a stance on a topic, did they verify before sending? What do they notice about their information consumption after observing their habits for a week, and what might they consider changing?
• Introduce students to the SIFT method (Stop, Investigate the source, Find better coverage, Trace claims, quotes and media to the original context). Note that some students may already know of this popular approach to addressing information online, so be sure to first ask if anyone can describe the method for others. Discuss how this approach may need to be modified in the age of AI. Challenge your students to design a modified method that accounts for the difficulty of finding a source and tracing claims where generative AI tools are involved.
• Given the difficulty or even impossibility of differentiating AI generated content from human generated content and tracing AI generated content to its source, help students focus on analyzing the content itself. Teach students lateral reading strategies and have them investigate claims in articles posted online using these strategies.
• Develop your students’ habit of asking questions by utilizing tools such as the Question Formulation Technique (registration is free) and the Ultimate Cheatsheet for Critical Thinking.

Tools for Shared Understanding

One of the most insidious consequences of AI generated disinformation is the way in which it can undermine our confidence in the reality of anything we see or hear. While it’s important that we prepare students to confront disinformation and to be aware of how generative AI will impact their information environment, we must also reinforce the importance of trust and shared understanding for the functioning of a healthy democracy.

• Help students recognize and overcome simplistic and dualistic thinking. Developing an awareness of the criteria and procedures used by different disciplines to verify claims will provide a framework for students to establish their ways of verifying claims. One approach might be to analyze the basis upon which generative AI tools such as ChatGPT makes claims.
• If confronted by a clear instance of mis- or disinformation in the context of a classroom or course-related interaction (for example, a student asserts the truth of a conspiracy theory that is blatantly false in a discussion board post), correct the inaccuracy as soon as possible. Point to established evidence for your claim. Help students see the difference between topics upon which we can engage in fruitful debate and topics where there is broad agreement, and to identify bad-faith approaches to argumentation.
• Ask students to create a healthy media diet for themselves. Where might they find verifiable information on topics of interest? What constitutes a good source of information on that topic?
• Promote empathy for others. We are more likely to believe inaccurate information about others if we are already predisposed to think of those individuals or groups negatively.
• Encourage students to see themselves as an actor within their information environment. Have them reflect on all of the sources of information they access and contribute to, including those within your class. Ask them to consider how they are using generative AI tools to inject content into that environment and what the implications of their decisions, and similar decisions of others may be on that information environment overall.

In the next installment of our series, we’ll dive a little deeper into the issue of bias and equity as it relates to AI. In the meantime, if you’d like to discuss digital literacy, artificial intelligence, or any other topic related to your teaching, please book an appointment for a one-on-one consultation with a member of the CAT staff.

References

Carolusa, Astrid, Yannik Augustin, André Markus, Carolin Wienrich.  Digital interaction literacy model – Conceptualizing competencies for literate interactions with voice-based AI systems.  Artificial intelligence, 2023, Vol.4, p.100114

Ecker, U. K., Lewandowsky, S., Cook, J., Schmid, P., Fazio, L. K., Brashier, N., … & Amazeen, M. A. (2022). The psychological drivers of misinformation belief and its resistance to correction. Nature Reviews Psychology, 1(1), 13-29.

Ferrara, E. (2023). Should chatgpt be biased? challenges and risks of bias in large language models. arXiv preprint arXiv:2304.03738.

Goldstein, J. A., Sastry, G., Musser, M., DiResta, R., Gentzel, M., & Sedova, K. (2023). Generative language models and automated influence operations: Emerging threats and potential mitigations. arXiv preprint arXiv:2301.04246.

Ng, Davy Tsz Kit, Leung, Jac Ka Lok, Chu, Samuel Kai Wah and Qiao, Maggie Shen. Conceptualizing AI literacy: An exploratory review. Computers and education. Artificial intelligence, 2021, Vol.2, p.100041

Organization for Economic Co-operation and Development, 2013

Wang, Bingcheng, Rau, Pei-Luen Patrick, Yuan, Tianyi. “Measuring user competence in using artificial intelligence: validity and reliability of artificial intelligence literacy scale.” Behaviour & information technology, 2022, Vol.42 (9), p.1324-1337

Wuyckens, Geraldine, Landry, Normand, and Fastrez, Pierre. Untangling media literacy, information literacy, and digital literacy: A systematic meta-review of core concepts in media education. Journal of Media Literacy Education, 14(1), 168-182, 2022  https://doi.org/10.23860/JMLE-2022-14-1-12

Dana Dawson serves as Associate Director of Teaching and Learning at Temple University’s Center for the Advancement of Teaching.

Jennifer Zaylea and Jeff Rients

In the previous two blog posts of this series, we addressed ways in which you may decide to incorporate AI into your classroom.  In this post, we offer suggestions for how you might creatively work outside the limits of AI by developing assignments and assessments that tools like ChatGPT cannot easily simulate.

Creating learning activities that prioritize human interaction (whether online or in-person), personal experience, and local knowledge can reduce the chance that your students will come to rely solely on AI-generative tools. This method has the benefit of prioritizing the humanity of the learning experience, because this method focuses student learning on the personal or community aspects of what we are teaching. It also can be more meaningful for students, motivating them to higher levels of effort.

Below are some possible assignments that are more difficult to simulate with generative AI tools. Remember, if you need assistance fleshing any of these ideas out or designing your own creative assignments, you can always schedule a one-on-one consultation with one of our staffers!

• Focus on human interactions in the form of attending events, arranging interviews, and conducting ethnographies of practitioners in your field. Ask for photos to be included as part of the student submission. Note: always include expectations for these kinds of activities up front in your syllabus, and keep in mind that it may be more difficult for some students to attend events and conduct in-person interviews because of work, transportation costs, etc. Having an alternative method for completing the assignment  handy in case of particular student hardship would be wise.
• Design assignments that ask students to meaningfully incorporate breaking news, local events, or niche references–forms of data or knowledge that AI tools have not been extensively trained on. Analysis of an event as it is still unfolding not only makes it harder for AIs to respond, but they also provide students with a much needed sense of the real connection of what they are learning with the world.
• Require “process papers” with follow-up on incomplete/incorrect steps to be revisited for iterative projects. For example, require that the second draft of a paper (or the second paper of the semester) must be accompanied by a memo outlining how earlier feedback was incorporated into the new submission.
• Ask students to record their making process for studio-oriented projects. This could take the form of a series of still images or videos, accompanied with a written reflection on the overall creative process.
• Strive to account for non-Western points of view outside the reach of most current AI datasets. Your students may need to use Google Translate to get the jist of some non-English websites, perhaps comparing them to similar English-language sources. Or they could interview people with non-Western perspectives.
• Use presentations, oral exams, classroom debates, and in-class writing activities to explore knowledge without immediate recourse to AI tools. Look for ways to connect these in-class activities to assignments outside of class, such as asking students to write a paper analyzing and evaluating a classroom debate.
• Consider using collaborative tools that allow for tracking of specific student contributions, such as Google Docs, Dropbox Paper, or Microsoft Word. A look at the edit history of the Google Doc for a collaborative project would allow you to see who was writing and who was revising the group paper.
• Utilize alternative assessments beyond high-stakes essays, such as group concept maps, gamification-based activities, or student podcasts.
• Assign multi-media composition work currently beyond AI capabilities, such as creating a zine that incorporates both words and images to document learning or designing a board game or card game that illustrates key course concepts in action.
• Ask students who are doing coding revisions, such as coding for graphic design, animation, computer programming, or websites to screen record their corrections while explaining why they made changes.  The assignment will encourage students to clearly articulate what the actions are accomplishing while reinforcing the skills they have learned.
• Have students write reflectively on how their past personal experiences intersect with their current learning.

Familiarity with a few of these approaches can also be helpful in the event that your otherwise AI-intensive course is disrupted by website crashes, fee structure changes, government regulation, or potential bans. A mix-and-match approach may best fit your learning environment, regardless. You may want one unit to focus on using AI tools and another that eschews them completely.

Just as with other teaching decisions (think absence or late work policies), your policies on the use of AI to complete work in your course will be a decision that you will have to make. As we’ve mentioned in earlier posts in this series, it will be important to include a statement on your syllabus, which you should discuss during your first class,  that explains your policies so that students understand clearly the expectations for your course, and so that the Office of Student Conduct has a standard by which to make decisions about any cases of academic misconduct. The CAT has developed sample syllabus statements that provide language you can adopt for your own use.

In the next installment of our series, we’ll look at how to foster students’ digital literacy skills in the age of generative AI.

Jennifer Zaylea and Jeff Rients both work at Temple’s Center for the Advancement of Teaching, where Jennifer serves as the Digital Media Specialist and Jeff Rients serves as Associate Director of Teaching & Learning Innovation.

Jonah Chambers & Jeff Rients

In the previous installment of this series, we outlined some ideas for how to put artificial intelligence to work as a tool in the classroom. This time, instead of adopting AI as a new educational technology, we propose that you use AI as an object of inquiry. A wide variety of controversial topics intersect with the rise of generative AI, providing a timely touchstone for the development of student critical thinking and media literacy skills.  Your students could productively engage with issues surrounding AI such as data privacy, dataset bias, environmental impact, intellectual property, and labor. Here are just a few examples:

• Collaboratively review the privacy policy and terms of use before using an AI tool. Will the students’ input become the property of the company owning the tool? What sort of privacy protections are in place?
• Contact multiple AI providers and ask them about their protocols for avoiding dataset bias issues like the now-infamous white Obama problem. Which responses suggest good corporate stewardship? How can their claims be tested?
• Compare and contrast the environmental impact of AI tools to the impact of search engine usage, cryptomining, the overall impact of the internet, etc.
• Debate the intellectual property concerns of large language AI tools versus other forms of intellectual appropriation.
• Attempt to map out which professions will be most impacted by AI in the short term. Which jobs will go away and which will be completely transformed?

Importantly, these sorts of investigations do not require you or the students to use the tools, making it an ideal choice for those who are uncertain about their willingness to join the AI revolution. However, some students may opt to interrogate the AI itself as part of their investigation. In this latter case, students might find it interesting to form their prompts as interview questions, tasking the AI with justifying its own existence!

However you want to handle AI in your course, it will be important to include a statement on your syllabus that explains your policies so that students understand the expectations for your course. The CAT has developed sample syllabus statements that provide language you can adopt for your own use.

Remember, although we are providing three different paths forward in the use of AI in your classroom, you don’t have to choose just one! A mix-and-match approach may best fit your learning environment. Your course might benefit, for example, from a unit critiquing AI followed by a unit that makes use of AI. Or vice versa.

In the next installment, we’ll look at creative assessments and assignments that not only discourage illicit student usage of AI but help your students learn as well!

Jonah Chambers is Senior Educational Technology Specialist at Temple University’s Center for the Advancement of Teaching. Jeff Rients serves as the Center’s Associate Director of Teaching and Learning Innovation.

Jonah Chambers & Jeff Rients

This week we’ll take a look at purposefully integrating the use of these AI tools into your course activities and assessments, particularly if the use of the tools can support student attainment of your learning goals. As with all educational technology, we generally recommend only those tools which support your previously identified goals And even when a tool does support your learning goals, consider the cognitive load of using the tool itself. Is the effort by the students to learn the tool worth the benefit gained? On the other hand, resisting the inevitable changes in our modern technological society is a well-trodden path to inconsequentiality. If you can, why not take advantage of these fabulous new AI tools to build more effective learning experiences for your students?

Here are some ways to make productive use of AI in your courses:

• Have students compete in creating the best prompt to elicit the most complete, useful, or interesting output to a course-related question or topic. Formulating a useful prompt requires clear articulation of the student’s own understanding, and comparing results allows students to practice their analytical skills.
• Ask students–on their own or in groups–to edit AI-generated texts. Editing could include fact checking, critiquing style, expanding upon, and adding references. Use collaborative documents (such as MS Word or Google Docs) to track changes to the original output. This process helps students develop critical thinking skills and digital literacy.
• Have students write annotated bibliographies based on guidelines you provide them and then compare those with AI-generated annotated bibliographies, noting where the AI-generated version produces errors or comes up short. This could work for other types of writing as well.
• Ask students to create a business plan, encouraging them to be as ambitious as possible and use the AI to provide ideas and feedback from a particular perspective.
• Task students with intentionally incorporating AI-generative tools into the writing process, such as to brainstorm paper topics, generate outlines, and proofread text they have written. You may want to check out these resources: Need an AI essay writer? Here’s how ChatGPT (and other chatbots) can help and How to… use ChatGPT to boost your writing.
• Share AI-generated responses with students to establish a new baseline of acceptable work, then expect better from them! Assign follow-up work that pushes student thinking further up Bloom’s Taxonomy.
• Allow students to submit an AI-generated first draft of a paper early in the semester, then focus their efforts on the revision process. Students submit both the AI draft and their revised version in a collaborative document (such as MS Word or Google Docs) so that you can track changes and see the version history.
• Per this paper, have students provide you with the list of the prompts they input and outputs they gathered in the process of writing a paper.

We’ve gathered some more specific activity and assignment ideas into this handy guide. Although any of these approaches will take some work to implement, this could be an opportunity for you to be an AI pioneer in your field, experimenting with possibilities these tools afford. You might even want to conduct a study to measure the effect of these tools on learning! Our resources on the Scholarship of Teaching and Learning can help you get started.

Keep in mind that you’ll have to decide to what extent and for what purposes students will be allowed to use AI, and then communicate those parameters clearly to students. That way students will understand clearly the expectations for your course. Key to any acceptable use of AI is transparency; you should insist that students identify any work that is AI generated. The CAT has developed sample syllabus statements that provide language for acceptable and unacceptable use of AI that can provide clarity to students and to the Office of Student Conduct for use in cases of academic misconduct.

Remember, although we are providing three different paths forward in the use of AI in your classroom, you don’t have to choose just one! A mix-and-match approach may best fit your learning environment.

In the next installment, we’ll look at how your students can learn valuable skills by integrating critical examination of generative AI tools themselves into your course activities and assessments.

Jonah Chambers is Senior Educational Technology Specialist at Temple’s Center for the Advancement of Teaching. Jeff Rients serves as the Center’s Associate Director of Teaching and Learning Innovation.

Dana Dawson, Ph.D

Generative AI tools like ChatGPT are already being used by students, and are likely to become ubiquitous in the workplace of the future, so ignoring them is not the ideal solution. We need to be actively thinking about how students are achieving the learning goals of our courses in light of their possible use of these tools, but also whether our courses prepare them for future careers in which they may be using AI tools regularly.

That being said, as with any other technology, we and our students need to approach the use of these tools critically. With regard to your teaching responsibilities, we see three approaches to the use of AI:

• Integrate the use of generative AI into your course activities and assessments;
• Integrate critical examination of generative AI tools themselves into your course activities and assessments; and
• Work around AI by designing assignments that are AI-resistant

If you’re not sure where to begin in deciding which of these approaches to take, we offer you this decision tree as a helpful tool. Determining how you will address generative AI in your classes will require you to reflect on your level of familiarity with the tools, take steps to become more familiar with them if you haven’t already, examine the learning goals for your course and determine whether AI can be useful in helping students to reach these goals,  and consider your readiness to carefully vet content students create with the help of or in response to the effects of generative AI. Finally, whatever your decision, you’ll need to speak with your students about the use of AI in your course.

In upcoming blog posts in this series, we will address strategies for the three approaches listed above as well as how to speak with your students about your approach. Note that you don’t have to pick just one of these approaches. We urge instructors to consider a mixed approach to incorporating AI in the classroom. In some cases, the use or analysis of content created by generative AI may help your students achieve a particular learning outcome and in other cases, it may be counterproductive. Revisit this decision tree not only at the outset of your class planning, but as you consider activities and assessments throughout the semester. Then articulate your decision to your students clearly on your syllabus. We have made syllabus guidance available to help you craft a coherent syllabus statement regarding use of AI in your courses.

Follow this blog series for continuing guidance on how to think about AI in your classes, and remember that you can make an appointment with a CAT developer to discuss your decision.

Dana Dawson is Associate Director of Teaching and Learning at Temple’s Center for the Advancement of Learning. Decision Tree graphic by the Center’s Graphic and Design Specialist, Emily Barber.

Emtinan Alqurashi, Ed.D., Jennifer Zaylea, MFA

Generative AI has revolutionized the way we interact with technology, opening new doors to faster and more efficient organization, and allowing more time for the creative and conceptual aspects of learning. However, there are considerable challenges that need to be addressed, especially when they could potentially impact how learning works in our classrooms. In our first blog of the AI series, we discussed what generative AI is and invited you to explore one of these tools (i.e., ChatGPT), in our second blog in this series, we explore some of the benefits and possible pitfalls of using generative AI in the student learning process. 

Some benefits of incorporating generative AI into the student learning process:  

• Access to large quantities of information. Generative AI is trained using a vast amount of information from a variety of sources. With this access to an enormous amount of information, it responds to a wide range of language-related tasks. Used productively, the greater access to information can potentially help students gain multiple perspectives on a topic, and spark inspiration that leads to greater creativity and ideation.  

• Speed of responding or processing. Generative AI provides quick responses to queries, prompts, and requests. This is particularly helpful when we need real-time responses or quick turnarounds for questions or requests. Rather than spending excessive time searching for information, students can gather information efficiently, and spend time focusing on comprehension and analysis of information.  Ultimately, this helps them advance to more complex thinking and learning. 
• Automation of routine tasks.  Because generative AI can produce human-like text, it can be used as a starting point for any type of content from drafting emails and articles to creating social media posts or course outlines. It can also be used as an editor to develop unstructured transcript or notes into a well-structured text. Students can save time and effort in the organizing and editing processes to allow them to focus on more creative and strategic aspects of their work. 

• Conversational manner of speech/writing. The human-like conversational ability makes AI an enjoyable discussion partner (but also provides serious concerns – see pitfalls below). In fact, generative AI can refine its outputs in response to your prompts in an ongoing discussion, helping you and your students to iteratively refine your thinking. For this reason, it can be very valuable to learn simple prompt engineering, which is the process of refining input instructions to achieve desired results. This can enhance the conversational capabilities of generative AI.   
• Assistive technology. Generative AI has the promise to improve inclusion for people with communication impairments, low literacy levels, and those who speak English as a second language. It can also improve ease of communication and comprehension in a variety of ways.   

Possible pitfalls when incorporating generative AI into the student learning process: 

• Engenders a false sense of trust. Generative AI’s conversational manner can be concerning as it can create a false sense of trust in the content it delivers. We must be aware that generative AI may disseminate inaccurate, biased, or incorrect information, and we must caution individuals against treating it as a source of truth.  
• Can generate inaccurate information. Generative AI has the ability to produce convincing responses to prompts and can offer properly formatted citations. However, it may not always provide accurate or reliable information as it has the ability to replicate outdated information, misinformation and conspiracy theories that exist in its data set. In all cases, the responses formulated by generative AI require careful vetting to ensure their accuracy. Therefore, it is important for students to take responsibility for verifying the accuracy and reliability of any information they obtain. This ensures students are not only learning correct information but also developing critical thinking and research skills that will serve them well in their academic and professional lives. 

• Can generate biased information. Just as generative AI can replicate inaccurate information, it can also replicate biased information. For example, if the training data used to train a generative AI contains biased or skewed information, the AI may inadvertently reproduce that bias in its results. Again, careful vetting of information is key to productive use of these tools. 
• Human authorship. Generative AI can produce responses that closely resemble pre-existing text, leading to an inability for the user to distinguish between AI-generated and human-authored text. Additionally, it is unclear which sources generative AI is using. This obfuscation of human authorship can make it challenging to attribute sources accurately. In fact, if you ask some generative AI tools to add references with citations and links, it can invent false sources, a phenomenon that is called hallucinating (although others are already becoming more sophisticated and can pull citations from real sources).   
• Simulated emotional responses. Generative AI can, in written form, simulate emotional intelligence, empathy, morality, compassion, and integrity. Devoid of nuance, AI simulates an emotional output by scouring the troves of text and training that has been provided to offer what seems most likely to be an accurate emotional response. It will be important for users of these tools to be cognizant of the fact that simulated emotional response is not the same as true interpersonal human connection and understanding. 
• Lacks context. When it comes to specific courses, class discussions, or more recent events, generative AI may not be able to establish connections between writing or arguments as they relate to the context of a course. 
• Proliferation of harmful responses. With prompt engineering, users can circumvent the safeguards put in place to prevent harmful, offensive, or inappropriate information from being proliferated in responses.  Additionally, there are concerns that misinformation (inaccurate) and disinformation (deliberately false) will be shared as accurate and reliable information. 
• Can potentially widen the education gap. The digital divide that can already exist between students with means and those without may be widened as more powerful AI tools are placed behind paywalls. In addition, gaps in digital literacy skills may exacerbate the widening of the education gap.  

While generative AI brings numerous benefits to students, the potential pitfalls must be considered. It is crucial for students to verify information, develop critical thinking skills, and exercise caution when relying on generative AI. These aspects are integral to developing digital literacy skills, which will be explored later in the series.  

In next week’s post, we will discuss how to decide about the use of AI in your class. If you have questions about generative AI and learning, book a consultation with a CAT specialist. 

Emtinan Alqurashi, Ed.D., serves as Assistant Director of Online and Digital Learning at Temple University’s Center for the Advancement of Teaching. 

Jennifer Zaylea, MFA, serves as the Digital Media Specialist at Temple University’s Center for the Advancement of Teaching. 

Emtinan Alqurashi & Jennifer Zaylea

Welcome to our summer 2023 weekly blog series A Survival Guide to AI and Teaching! In today’s rapidly evolving world of technology, one innovation that is altering higher education is generative AI (such as ChatGPT, Dall-E, Bard, Bing, or visit SourceForge for their 2023 Best of listing). The impact of advanced generative AI on higher education assignments and assessments is complex. Our goal in this series is to help you understand and proactively address student use of generative AI at Temple.  Over the coming weeks, we will provide an overview of how generative AI works, discuss some benefits and pitfalls of using these tools, help you decide whether you should integrate generative AI into your instruction, and offer supportive suggestions for how to whatever your decision is. We will address the topics of digital literacy, academic integrity and how to talk to your students about ChatGPT and similar tools. Join us as we delve into the influence of AI on the future of higher education and the ethical considerations it engenders for students and faculty. In this first blog post in the series, we will explore the question of what generative AI is and how it works. 

What is generative AI? 

Generative AI (such as ChatGPT, Bard, Bing, Dall-E, etc.) is a type of artificial intelligence that has the capability to create surprisingly coherent text and images. They are large language models that can write and converse with users by drawing on an enormous corpus of text from a variety of sources—including books, web texts, Wikipedia, articles, internet forums, and more—on which they have been trained. These models are growing progressively larger: for instance, GPT-4, the most recent iteration of GPT-3, has 170 trillion parameters compared to its predecessor’s 175 billion. Generative AI does not have cognition; that is, it can’t think. Instead, in a similar manner to the autocomplete function that works in other applications you use every day, it works by finding and replicating the most common patterns in speech available in its dataset. Note, however, that AI technology is constantly evolving and improving, and we cannot be sure what its future capabilities will be. For an expanded explanation of how these technologies work, visit this article “A curious person’s guide to artificial intelligence.” 

How does AI work?  

Generative AI have been developed to provide a detailed response, in a conversational way, to a given prompt in the form of a question, request, or instruction. Rather than providing the user a listing of information sources generated by a prompt, as search engines like Google do, it structures the response in a more humanistic question and answer format.  

We recommend you spend time with a generative AI tool to understand what it produces. Not only will this reduce any anxiety you may be feeling about this unfamiliar technology, but it may help you discover unrealized uses for generative AI in your courses and your discipline. Using prompts from your own classes will give you a good idea of what students can generate using the tool. We’ll lead you through the steps for interacting with one of the AI tools, ChatGPT, which is available by visiting OpenAI’s website 

Let’s try it out! 

Step 1: Visit OpenAI and create a free account, by clicking on the TRY CHATGPT button, and then clicking on the SIGN-UP button.   

Step 2: Become familiar with the interface.  Navigation tool descriptions are provided below the image. 

Step 3: Use the text entry box (area 8, “Send a message…”) at the bottom of the screen to type a prompt. This can be a question or a specific request. Press the paper airplane icon to submit.  We have provided a video example, Prompting In ChatGPT, showing how to do this. 

Step 4: ChatGPT will “type” out the response in real time. When it’s done you can click the REGENERATE THE RESPONSE button, if you want to see different wording of the output, and/or provide feedback regarding the accuracy of the response by clicking the THUMBS UP and THUMBS DOWN buttons. This will help teach the AI how to work better with you. 

* Note: Each prompt kicks off a conversation. You can enter follow-up prompts or change the subject entirely. It will remember previous conversations. 

Step 5: If you don’t think the response was good enough, you can simply ask it to try again and provide new parameters or details to guide its response. 

Step 6: You can also let ChatGPT know when it’s incorrect about something.  

Once you have explored ChatGPT, or other generative AI tools, you will have a stronger understanding of how the tools work and how you will have more confidence when addressing them in classes, determining how you might use them in assessments, or even think about using them to perform faculty administrative tasks. 

In next week’s post, we will delve into the potential benefits and pitfalls of generative AI for students and educators. If you have questions about generative AI and its application in your classes, feel free to book a consultation with a CAT specialist. 

Emtinan Alqurashi, Ed.D., serves as Assistant Director of Online and Digital Learning at Temple University’s Center for the Advancement of Teaching. 

Jennifer Zaylea, MFA, serves as the Digital Media Specialist at Temple University’s Center for the Advancement of Teaching.