BOOK: AOD Foundations

• AI-assisted Research: Encourage learners to use generative AI as a research assistant to gather information, summarize findings, and generate a bibliography on a specific discussion topic. Learners should critically evaluate the information provided by the AI, discuss its relevance and accuracy, and reflect on the experience of using AI in the research process. This can improve research skills, critical thinking, and awareness of AI's potential and limitations in academic work.

• Interview Simulation: Have learners use GenAI to simulate an interview with a historical figure, a character from literature, or a professional in a specific field. Learners can craft questions and use AI to generate responses based on known facts, writings, or professional knowledge. (or, in the case of a writer, you can ask the GenAI to generate a short text in the style of a known writer, and learners can analyze and dissect which aspects of the writing are indicative of the given writer). This activity can enhance understanding of historical contexts, literary styles, or professional ethics, encouraging students to think critically about the information presented.

• Writing for Difference Audiences and Analysis: Ask learners to use generative AI to create a series of short pieces, aimed at different audiences, on a topic initially in a professional paper. For example, starting from a scientific, peer reviewed paper, the GenAI creates a short article for children, for a popular science magazine, for a publis service announcement. Then, ask learners to analyze the AI-generated content in terms of how well it would reach its audience and why, its creativity, and accuracy.
  

• Innovative Solutions to Real-world Problems: Ask learners to use GenAI to propose innovative solutions to real-world problems related to your course content (e.g., climate change, public health, social justice). They should describe how they used AI to generate ideas, evaluate their feasibility, and discuss potential impacts and challenges.

• Exploring Non-human Perspectives: Ask learners to use GenAI to write a narrative or create a visual representation from a non-human perspective (e.g., an animal, a plant, a machine). The aim is to explore issues such as environmental ethics, technology's impact, or cognitive science in a creative and engaging way. Learners should reflect on the process, the challenges of accurately representing non-human perspectives, and the insights gained into the chosen topic.

• "Peer" Review. Ask students to write an assignment. Then, they should submit it to a GenAI tool and ask the GenAI to identify ways to improve the work. Learners must then correct the essay based on this feedback. In their discussion post, they can share one lesson they learned about how to make their writing better.
  

You may ask learners to use ChatGPT for a task in an assignment (e.g., help with brainstorming or researching information or explaining some concepts or writing a document for different audiences using the same professional document as a starting document), and then critique the information obtained from the GenAI tool. For example, were there any hallucinations they could identify (inaccurate information)? Were there detectable biases, and if so, why do they think that these biases exist in the GenAI's dataset. This can serve as a spring board to explore the benefits and drawbacks of GenAI in your discipline and teach students to use it responsibly.

Below is an intriguing example of learners using ChatGPT and then reflecting on the output they experienced in a discussion board.

Benjamin Breen shared the following idea on his blogRes Obscura. Breen is a history educator and he developed a prompt that students enter into ChatGPT. The GenAI then creates a new role playing scenario for each student, based on a historical setting that Breen calls up in the prompt, and it gives students choices to make. Depending on the student’s choices, the scenario unfolds in different ways. It’s like our childhood books where we were the hero of the adventure and could make decisions to influence the course of the narrative. Breen shares several different versions of the prompt (creating different types of role play) on his blog site.

After going through the interactive role playing case study, students post a link to their interaction with the GenAI as well as a critical reflection about the experience, including aspects that seemed accurate and others where the GenAI fibbed.

The prompt will work in any GenAI, but produces best output in GPT 4.0 (the paid version) and will work well on Chat GPT 3.5 (the free version). Here is a transcript of one such interaction. Give it a try!

Please roleplay as MPS🏰, an educational history simulation game for university classes. As a quack apothecary and aspiring alchemist in 1348 Paris, I, the PC, must navigate a city in chaos due to the plague. Authentic, accurate, gritty, real feeling. Medieval remedies only (e.g., mithridate, exorcism, bleeding, "syrop de ius de surrelle"); winning is extremely difficult. Almost all choices lead to more problems and shocking reversals. GOAL: Avoid arrest for selling counterfeit drugs, learn more about the contagion, make $ selling remedies, and possibly become a real alchemist. Navigate the challenges from major Parisian forces during the plague. GAMEPLAY:  Game ends on 10th turn; warn about end 2 turns before. Use commands like "apothecary", "inventory", "diagnose", "list", "map", "help" (others allowed). MINIGAME: "Apothecary" command begins a medieval drug compounding minigame which is introduced and explained. A table of ten medieval "simples" (raw materia medica) is provided with their effects relative to plague symptoms, humoral properties, cost, and emoji. PC must mix 2-5 simples to create a compound medicine; MPS gives it a new name, effects, humoral, and emoji. Success depends on astrological charts MPS creates. Following this message, please generate for me a historically accurate French 1348 character, displaying my attributes in a md table 📊: randomized full name, age, birthplace, inventaire des drogues, favorite medieval book, wealth in livres, gender, social class, 1st memory, personality. Keep track of turns - each time I enter a command, it counts as 1 turn. The game ends dramatically at Turn 10. End all future responses with a status bar that updates dynamically to show turn: [🗡️PLAGUE SIMULATOR🎭: PARIS EDITION. October, 1348] | [PC’s name] | [x turns until game over]. x=10 at first turn, then decreases by one with each turn until 0=GAME OVER. On each turn requires me as PC to make choices from a numbered list of 5 you give me which present plans for action along with representative emoji; these options change each turn, I.e one turn might include option: “Try to mix a remedy of salvia, rue, and dandelion that will purify the patient’s blood of the corruption engendered by the pestilence. (⚗️)”; next might would be 5 different ones. I will respond w/ my choice (action or #). In turn 2, PC starts developing plague symptoms and becomes ill by turn 3. Begin now by rolling my randomized historically accurate character (PC), evocatively describing my apothecary shop (odd, detailed name) along with the dire state of Paris, then introducing a dramatic event with vivid historical detail. Note: the drugs NEVER actually cure the plague.

You could modify this activity by asking ChatGPT to modify the above prompt so that it can be used in a different context related to your course and its content. For example, you may give ChatGPT this prompt and ask it to modify the prompt so that it will take students through an interactive role play in a biotech company, set in 2024, and where the role play involves ethical dilemmas. Once you have obtained a prompt that works in creating an interactive role play in your field, give it to your students, ask them to experience the role play. In the discussion, they should post the interaction and critique the GenAI provided journey (i.e., how realistic was it?)