Algorithmic Education

I have posts - both present and forthcoming - regarding the complex interplay of ethical responsibilities, assignment quality, and even holistic educational construction when artificial intelligence meets pedagogy. It's a topic that is both fascinating and prescient, not least of which because I am, of course, a professor deeply embedded with artificial intelligence. For today's offering, however, I'd like to forgo the complex, rigorous debate, and instead offer some "boots on the ground," actual augmented assignments that have appeared in various forms within my classes. I encourage you to take these, modify them according to your content (and students') needs, and see where an augmented classroom can get you.

First, a few notes:

• Don't overprepare or offer excess time for students to use the tools. In addition to having the lowest learning curve of nearly any emergent technology, AI also offers an incredible boost to efficiency on the part of student work - make use of it! This isn't to say that no discussion is required - you always want to include contexuality, as well as best practices, like prompt engineering - but the need to explain detailed histories and methodologies, as well as hard skills, is far less demanding than other software. There is no "learn Photoshop" before learning design, or "learn Excel" before learning data science, for example.
• Most of these assignment types appear in my Digital Humanities classroom, but others appear, at various intervals, in Game Studies, AI Culture (expected), Tools and Technologies, or other areas. I've attempted to note which classes make use of these assignments, but also want to note that with only a few changes, these assignment types may work in nearly any setting, with any content area. Remember that, at the end of the day, definitionally, AI is automated, trainable communication. Nearly every assignment under the sun may benefit in some way.
• It should go without saying, but I'll say it anyway: make sure your chosen application is moral, ethical, legal, and foundational. Don't use AI (or encourage students to use) AI in immoral ways (deepfakes), unethical ways (plagiarism or its AI twins), illegal ways (track governance at every level, from federal to institutional codes of conduct), or un-foundational (don't employ AI if the result detracts from the trustworthiness or integrity of the result, either of content or researcher).
• Finally, remember that you may choose to deploy these activities as accommodations rather than standards. Some students, depending upon accommodation measures, may benefit significantly, while others will only exploit the power AI offers. It's incumbent upon us to solve that case-by-case equation.
• Glossary - MDAP Metadata-Defined Agent Persona (My proprietary prompting methodology for LLM Agents); MPE Multiple-Persona Embodiment (Another proprietary prompting methodology); RAG: (Retrieval-Augmented Generation); DH (Digital Humanities); DS (Data Science); AI (AI Culture); IVGS (Intro to [Video] Game Studies); T&T (Tools and Technologies)

SECTION ONE: BASIC AUGMENTATION

Research Assistance (location/confirmation)

Used In: Digital Humanities (DH), AI Culture (AI), Intro to Game Studies (IVGS), Tools and Technology (T&T)

Use Case: Expediency of research, efficiency of source location, source vetting and confirmation

Tools: GPT+, specifically the Consensus GPT, or a custom-made GPT (used only in AI Culture)

Abstract: A simple research assistance tool used to augment Google Scholar and Google Books searches. Helps students identify, locate, vet, and examine research articles and authors. Best deployed in the opening phases of research, and does a far more focused job of suggesting and collecting articles for students to read than does a mere Google search.

Method: Ask students to compare their research to that from Consensus (after explaining what a hallucination is, of course).

Research Assistance (summary/output)

Used In: Digital Humanities (DH), AI Culture (AI), Intro to Game Studies (IVGS), Tools and Technology (T&T)

Use Case: Expediency of research, source vetting and confirmation

Tools: GPT+, Claude

Abstract: A research assistance tool that makes research more efficient and does a better job of summarizing articles than do research topic abstracts. Often pairs well with a lesson on skim reading, the importance of topic sentences, and/or the importance of close reading.

Method: Use a basic, unprogrammed chat (can deploy a full MDAP for this if you prefer) to summarize lengthy or complex research articles, translate jargon, or simply identify which articles are prescient to research so the student can follow up with a close read. Demonstrate that while efficient, the summary does not and can not replace a quality close-reading methodology. Note: your students need to know what quality close reading looks like. I recommend pairing this (as I do in DH) with a discussion of Nicholas Carr's "Is Google Making Us Stupid?", Jaron Lanier's "Digital Maoism", and/or Katherine Hayles' "How We Read: Close, Hyper, Machine".

Content Ideation

Used In: Digital Humanities (DH), AI Culture (AI), Intro to Game Studies (IVGS), Tools and Technology (T&T)

Use Case: Writer's block and/or choice paralysis

Tools: GPT+

Abstract: A basic deployment of an ideated chat - best deployed with a functional MDAP - which assists students discover and focus their topics for ultimate research and construction or writing. May be deployed for everything from essays to PBL. I build and share my own MDAP agent for this use case, since I want to mitigate hallucinatory text as much as possible.

Method: Deploy a standard agent so that students may craft ideas, whittle them into more focused patterns, or break writer's block. Very useful in creative writing, the arts, or when just embarking on a large-scale research project.

SECTION TWO: AUGMENTED CLASSROOMS

Socratic Tutors

Used In: Digital Humanities (DH), to Game Studies (IVGS), Data Science (DS)

Use Case: Study aid

Tools: GPT+, best deployed with verbal commands and RAG

Abstract: A basic deployment of an ideated chat - best deployed with a functional MDAP - which converses with students and continues a conversation about any topic.

Method: Deploy a custom MDAP with RAG - loaded with a pdf of your class' chosen text(s), and instruct it to converse in never-ending form with students about any content you wish. This conversation is particularly helpful for students requiring accommodations, and can help drive repetition in study hours without man hours (beyond building the initial agent).

Socratic Dialogue

Used In: Digital Humanities (DH), to Game Studies (IVGS), Data Science (DS)

Use Case: Study aid

Tools: GPT+, best deployed with verbal commands and RAG

Abstract: A secondary deployment of the Socratic tutor, using MPE (Multiple Persona Embodiment) instead of MDAP to create a functional conversation "room" wherein students may participate as a member of a group.

Method: Deploy a custom MPE loaded with differing perspectives of your student's material, such that you can explore controversial or difficult/challenging subjects from multiple perspectives simultaneously.

Role-Play

Used In: Digital Humanities (DH), AI Culture (AI), Intro to Game Studies (IVGS), Tools and Technology (T&T)

Use Case: "In-person" practice

Tools: GPT+, best deployed with verbal commands

Abstract: A custom MDAP equipped with OCEAN (or, in light of the personality matrix, Advanced Voice if you prefer) which acts as instructed for training purposes. May be used to deliver feedback, argue with students to test rhetoric or research, practice interpersonal interactions, or even public speaking.

Method: Key here is the personality of the Agent. I personally prefer an MDAP with OCEAN (that is, an agent with both a job and a persona) which can act in specific ways. Train students to deliver news to a patient with an appropriate bedside manner; deliver constructive feedback to peers (especially for those looking to get into management or HR); or engage in rhetorical training or public speaking. Assign a persona to each agent to increase the level of difficulty. I use the OCEAN (Big 5) methodology to acquire pinpoint control over how, for example, neurotic, I want the agent to ultimately be. Note: If re-using an agent in this context, use a new chat for each student so you don't overrun the context window.

Creative Bounds

Used In: Digital Humanities (DH), AI Culture (AI), Intro to Game Studies (IVGS), Tools and Technology (T&T)

Use Case: Subject matter construction

Tools: Midjourney, Udio, or Runway (or any other multimedia AI)

Abstract: Illustrated texts, or textual illustrations. Use AI to create image, audio, or video accompaniment to student work (illustrations for everything from presentations to web sites to creative fiction/nonfiction); or, create the multimedia object first and ask students to write/react/discuss it.

Method: Option 1: Ask students to use multimedia AI to augment an existing assignment. Option 2: Create a multimedia object and use it as the basis for an assignment. Write a story about picture x or video y; Compose a song about topic x using your research; add multimedia and/or visual rhetoric to your project (website, newspaper layout, etc).

Design Principles and Visual Rhetoric

Used In: Digital Humanities (DH), AI Culture (AI), Intro to Game Studies (IVGS), Tools and Technology (T&T)

Use Case: Efficiency and theory

Tools: Midjourney, GAMMA, or Runway (or any other multimedia AI)

Abstract: Use of AI to create artifacts with which students may engage.

Method: Ask students to use proper design principles to "grade" an AI-output website design (GAMMA, Midjourney). Use AI to create visual rhetoric in other assignments (like presentations). Ask for AI-gen videos instead of PowerPoints, or use AI to create (and then modify) the PowerPoints. Discuss visual rhetoric within AI-generated content.

Coding Artifacts

Used In: Digital Humanities (DH), AI Culture (AI), Intro to Game Studies (IVGS), Tools and Technology (T&T)

Use Case: Deployment and engagement

Tools: Claude, GPT+ with Canvas (and a chosen IDE)

Abstract: Use of AI to create interactive content

Method: This lesson has a special place in my IVGS classroom, where we engage literary principles with game texts; however, one assignment has students create a game text of their choice which queries the human condition through literary principles and brings it to life using Claude's artifacts. This allows us to examine both the literary principles of the text as well as Juul's ludic principles simultaneously. Strongly recommend pairing this with Jesper Juul's Half-Real (my IVGS textbook) or Mark Rosewater's "20 years, 20 lessons", an excellent, and jargon-free, introduction to game design. This assignment also finds a home in T&T, where students create a working prototype for a Shark-Tank style pitch to investors and both DH and AI, where I use it to automate the PBL objectives in major graded assignments (called "Digital Objects" and "AI Objects," respectively).

Data Analysis

Used In: Data Science (DS)

Use Case: Efficiency and Accuracy

Tools: Wolfram, or GPT+ (Wolfram GPT)

Abstract: Use of AI to examine statistical output

Method: This class, Humanities Data Science (HDS), does not focus on strict data science principles (for example, we don't discuss machine learning archetypes and how they may be deployed), but rather Humanities applications of automated trends, like marketing, multimedia creation (music producers, etc.), art, design, and literature. Therefore, it isn't incumbent upon this class to know the hard principles and statistical analyses behind the data, rather how we can deploy and ultimately act upon their output. Rather than provide examples and ask for analysis, however, I ask my students to engage in the entire data pipeline, from scraping to sourcing to analyzing to creating an analytics record to statistical analysis, all by hand. To assist with this process - and since statistics isn't a class prerequisite - we deploy Wolfram to assist with complex math like linear regression and Bayesian analysis.

SECTION THREE: FULL AUGMENTATION

Gamification

Used In: IVGS

Use Case: Ludic Principles

Tools: GPT+, or a GPT with specific coding skill (there are many) as well as the IDE of your choice

Abstract: Gamify a lesson

Method: Rather than asking students to create a game text, make one of your own. Use an LLM to write and proof the code, your IDE of choice to compile, and (usually) a website to deploy a game text that has its features rooted in your topic of choice. Want students to learn rhetoric? Make it into a game, perhaps one where they converse on a debate stage, for example. The sky is the limit. These interactive texts work even better when augmented with other multimedia AI, like Midjourney images, Runway videos and voice, and GAMMA website menus.

Cryptography

Used In: DH

Use Case: Investigative Research

Tools: CryptoCrack, CrypTool

Abstract: Holistic Research using AI to uncover data

Method: The final project in my DH classes - and one of my favorites - has groups of students choose an unsolved historical cipher text, of which there are many, and use Neural Cipher identifiers available from the above tools to identify the cipher type. This is then paired with a Brute-force cracking tool, as well as historical and biographical research, to engage students with real-world forensic examination on everything from religious texts (Voynich) to pirate treasure (La Buse) to criminal justice (Zodiac).

I have more which I'll share in a later article - hopefully this gets the ball rolling on some ideas of your own, which I'd love to hear about!