TEACHING WITH AI Slides with Citations

• DELTA COLLEGE (Michigan) Workshop & Keynote
• SUSQUEHANNA UNIVERSITY Workshop
• GATEWAY TECHNICAL COLLEGE (Kenosha) Keynote and Workshop
• MONTANA TECH: Morning Workshop, Afternoon Intro & Afternoon Workshop (Teaching with AI)
• NEIU: Keynote & Workshop
• Workshop A Part 1 Tools.
• Workshop A Part 2 Teaching.
• Workshop B (shorter but complete).
• Complete AI Slides with Citations (143 Pages updated July 25, 2026)
• New Jersey County Vocational-Technical Schools AI Strategy Slides and Citations (PDF August 6, 2026)
• Slides and Citations for Re-Designing Learning in the Age of AI.
• AI Literacy Handout  (2-pages front and back)
• AI Strategy Slides and Citations (60 Pages)
• Faculty Discussion Questions for the 2nd Edition by Gemini Notebook
• More Handouts and Templates

ASSIGNMENTS and PROMPTS

• AI Pedagogy Project at Harvard
• The free Pedagogical Promptbook by David Wiley is an excellent open source set of tested and verified prompts for teaching and assignment design.
• ADAPT (formerly “Understanding AI Use Through Assignment Design”), with an Ethical Matrix for AI use, from Penn State Behrend
• Here is an OER Gemini Notebook from Tom Haymes loaded with source material about AI and Pedagogy.
• Harvard Business Impact page on Teaching with AI. They also have a good list of short articles on AI.
• Great set of general resources (with lots of useful prompts): https://www.aiforeducation.io/
• Prompts from Levy and Albertos for improving classes and assignments.
• 22 Ideas for New Assignments from Steven Mintz
• Here is a great way to get feedback on any assignment or lesson plan from Doan Winkel.
• The Course AI Resilience Tracker (CART) Tool from Oregon State University will help you evaluate and enhance the resiliency of your course in the context of generative artificial intelligence tools. You enter information about your learning outcomes, students, materials and activities etc. and then get feedback on how to make your course more resistant to AI. (OSU does not collect or track any information that you input. Closing, reopening, or refreshing the page will clear all entered information.)
• Lance Eaton’s Prompt Library for a wide variety of faculty tasks
• Harvard Business Publishing
• UCF Teaching Repository for AI-Infused Learning
• UCF: Open source book with AI activity prompts (2023)
• MLA AI Pedagogy Community
• Anthropic’s prompting best practices (the old Claude prompt library was retired)  https://platform.claude.com/docs/en/build-with-claude/prompt-engineering/claude-prompting-best-practices
• Assignments that Include Text Generators from Anna Mills
• More free teaching tools
• Jeanne Beatrix Law (from Kennesaw State University) calls her “authentic first, technical second” approach “rhetorical prompting.” Students start by generating, reflecting and refining ideas before they even settle on a thesis. Here is her custom GPT support bot that uses OER resources: https://chatgpt.com/g/g-KwpWcnhqe-openstax-writing-guide-assistant
• Anna Mills Writing and Research Prompts at MyEssayFeedback
• Papyrus AI Prompts for Writing and Research
• Ethan Mollick Prompts: https://www.moreusefulthings.com/prompts
• AI-Integrated Assignments from Derek Bruff at UVA
• A prompt for creating engaging handouts from Jason Tangen: more of his tools and prompts here
• Prompts for teaching from Cynthia Alby (How to create a case study and much more.)
• Harvard GenAI Library for Teaching and Learning
• Cornell AI in Assignment Design
• Duke AI in Assignment Design
• A Student Guide to Navigating College in the Artificial Intelligence Era 
• 10 Best Practices for AI Assignments in Higher Ed
• Revised Bloom’s Taxonomy that includes generative AI created by Oregon State University
• A prompt library for simulations and career support for students in technical fields. (Over 300 prompts broken down by subject areas: simulated interviews, problem-solving scenarios, work-based simulations, and more.)

PRIVACY

Last verified: July 27, 2026. Everything below describes consumer and free accounts. If your campus has an institutional agreement (like Copilot) the terms are usually much better: training is normally off by default and retention is contractual. The single most common privacy mistake on campus is assuming the enterprise terms apply when faculty and students are actually signed in to the free consumer version.

• Your campus needs a clear and simple guide for everyone that lists what tools you can use and what information you can safely enter into each. Here is a great example from Tufts. Note how they clarify the types of data: public, confidential and restricted (which includes FERPA data).
• Whatever your campus policy says, the practical questions for any given tool are always the same four: Is my conversation used to train the model? Can I turn that off? How long is it kept? Who else sees it? The table below answers those four questions for the major consumer tools.

WHAT THE PLATFORMS SAY THEY DO (consumer and free accounts, checked July 2026)

Tool	Chats used to train by default?	Can you turn it off?	Retention	Worth knowing
ChatGPT (OpenAI)	Yes	Yes. Settings > Data Controls > “Improve the model for everyone”	Deleted chats and Temporary Chats are removed within 30 days	OpenAI began testing ads in ChatGPT in the US on February 9, 2026, on the Free and Go plans. Personalized ads can draw on past chats and memory. Plus, Pro, Business, Enterprise and Edu accounts are ad-free. Advertisers do not receive your chats. This matters for students, who are mostly on the free tier.
Claude (Anthropic)	Yes, since August 2025	Yes. Privacy Settings > Model Improvement. Incognito chats are always excluded.	5 years if training is on, 30 days if off	Conversations flagged for safety review can be used regardless of your setting, and Anthropic does not say what triggers a flag. Flagged inputs are kept up to 2 years and safety scores up to 7.
Gemini (Google)	Yes	Yes. Turn off “Keep Activity” or use Temporary Chats.	72 hours when Keep Activity is off, otherwise saved to your Google account	When Keep Activity is on, a sample of conversations may be read by human reviewers. This is a change from 2025, when Gemini offered no training opt-out.
Copilot (Microsoft)	Yes	Yes. Privacy > Training on conversation activity	Conversation history kept 18 months. You can delete individual items or the whole history.	Opting out of model training does not exclude your conversations from other product improvement, advertising, safety, security or compliance uses. Microsoft serves personalized ads in Copilot, and your conversation history can feed them unless you turn personalization off. All of this applies to personal accounts. A work or school login falls under Enterprise Data Protection instead.
Meta AI	Yes	No opt-out in most regions	Not clearly specified	Since December 16, 2025, Meta uses AI chat content to target ads across Facebook, Instagram, WhatsApp and Messenger, with no way to opt out. Religion, health, politics and sexual orientation are excluded from targeting but may still be stored. The EU, UK and South Korea are carved out.
Le Chat (Mistral)	Yes on the Free, Pro and Education plans	Yes. “Allow your interactions to be used to train our models” in the Admin Console, or Settings > Data & Account Controls on mobile. Teams and Enterprise plans are opted out by default.	See Mistral’s privacy documentation. Zero Data Retention is available on some plans.	Uploaded documents count as input data, so an instructor on the Education plan who uploads student work is sharing it unless they opt out. Mistral collects less data than the US platforms and is covered by GDPR, which is much of why it topped the 2025 ranking below.
Grok (xAI)	Yes	Yes, in account settings (not independently verified in July 2026)	See xAI’s privacy policy	In August 2025 roughly 370,000 shared Grok conversations were indexed by Google, including medical questions, uploaded files and at least one password, because the share links carried no “noindex” instruction. xAI fixed it after reporters found it. A strong privacy policy is not the same thing as a private product.
DeepSeek	Yes	No consumer opt-out documented!?!	Not clearly specified	Data is stored on servers in China under its stated privacy policy, and the policy itself is unusually thin. Several governments have restricted it on official devices.

• The best current source is the 2025 Foundation Model Transparency Index from Stanford’s Center for Research on Foundation Models (December 2025). This is the third annual edition, and the news is bad: the average score fell from 58 in 2024 to about 40 in 2025, a 17-point drop in a single year. Companies are most opaque about training data, training compute, and what happens to your data after deployment. DeepSeek and xAI were scored for the first time. The full paper is here. Note that this measures what developers disclose, not what actually happens to your data, but it is academic work with a public rubric and it is current.
• Tran and colleagues surveyed 300 US ChatGPT users and found that 82% rated chatbot conversations as sensitive or highly sensitive, higher than they rated email or social media posts. Nearly half had discussed health topics with ChatGPT anyway, and over a third had discussed personal finances. What changed people’s judgment about sharing was procedural (consent, anonymization, removing identifiers) and not who received the data or why. This is a preprint with a US-only sample of 300.
• The Incogni Gen AI and LLM Data Privacy Ranking is widely cited and still worth reading, but it is a dated snapshot. The data was collected May 25-27, 2025 and the article has not been substantively revised since June 30, 2025, despite the “2026” that now appears in the page title. Incogni is also a paid data-removal subscription service owned by Surfshark, so this is vendor research, although the methodology is disclosed and the full dataset is public).
  • Its 2025 findings were: Le Chat least privacy-invasive, followed by ChatGPT and Grok; Meta AI worst, then Gemini and Copilot, with DeepSeek also among the worst; ChatGPT most transparent about training; and every platform collecting from “publicly accessible sources.”
  • Three of those findings no longer hold. Gemini now offers a training opt-out and Temporary Chats, so it does not belong on the “no opt-out” list. Anthropic no longer claims that user inputs are never used for training, which changed in August 2025. And ChatGPT’s free tier now carries ads that can be personalized using past chats.
  • Note too that this study scores privacy documents, not actual data flows. It could not have caught the Grok indexing failure, because xAI’s policy was fine and the engineering was not. It could not anticipate the Meta or OpenAI advertising reversals. Incogni also used California-specific policies where available, so the results may not describe what a user elsewhere actually experiences.
• European Data Protection Board Report (April 2025) on AI Privacy Risks & Mitigations – Large Language Models (LLMs). This is a regulatory framework rather than a leaderboard, which is why it has aged much better than the rankings have.
• AI Transparency Rankings from Americans for Responsible Innovation (January 2025). This ranks seven frontier models on 21 documentation metrics. It is now more than a year old and largely superseded by the Stanford index above.
• Here is a massive spreadsheet by Sarah Wood that lists most of the AI models and API tools and scores them for a variety of privacy and environmental issues. It scores them for appropriate use for elementary, middle and high school, but very useful for higher education too.
• The Common Sense Privacy Program evaluates edtech and AI products against a consistent, published question set. It is built for K-12 rather than higher education, and like the rankings above it evaluates policy documents rather than actual data flows. But the rubric is transparent, and their evaluation questions are a good template if you are assessing a tool your campus is considering.

The ENVIRONMENT

• Jon Ippolito’s App “What Uses More” allows you to compare the environmental footprint of digital tasks.
• Andy Masley has written extensively about this and his blog provides detailed resources. I still think his summary about why using ChatGPT is not bad for the environment is a good place to start.
• Exploring the Intersection of AI & Environmental Impact is a curated collection of knowledge, tools, and insights at the intersection of artificial intelligence and sustainability.
• From the Penn State Institute of Energy and the Environment: Why AI uses so much energy—and what we can do about it
• Energy and AI from the International Energy Agency (April 2025)
• MIT Technology Report (May 2025): We did the math on AI’s energy footprint. Here’s the story you haven’t heard: The emissions from individual AI text, image, and video queries seem small—until you add up what the industry isn’t tracking and consider where it’s heading next.
• Andy Masley’s Response to MIT Report.
• Epoch AI is a reliable third-party source of info on AI, including the environment.
• The Sustainable Agency list of estimates.

AI POLICIES

• Anthropic now watermarks Claude’s text (August 2026). This is not a detector. It changes which of several equally good words Claude picks, using a key, so anyone holding the key can estimate whether Claude was involved. Read the limits before you lean on it: it cannot tell “Claude wrote this” from “Claude heavily edited this,” it barely registers on proofreading or on short passages, a full rewrite removes it, and the detection API is not out yet. Other companies signed the same EU code and will ship their own.
• United University Professions, representing 42,000 SUNY faculty and staff, ratified a five-year contract in July 2026 saying that courses stay under the “direction and responsibility” of humans and that humans hold “ultimate accountability” for work done by AI. It passed with 97.7%. Almost every AI policy you will read was written by an administration. This one was bargained.
• The EU AI Act became applicable on 2 August 2026, and the transparency rules are now in force, which is why Claude started watermarking. But the high-risk rules that actually cover education — admissions, scoring exams, evaluating learning outcomes — were pushed to 2 December 2027 by the AI Omnibus. A lot of coverage still says August 2026. The obligations did not change; only the date did.
• Resources and AI policies in your course Stanford syllabus sample language.
• Menus, not traffic lights: A different way to think about AI and assessments by Danny Liu.
• A great set of statements, sample syllabi statements and more from Ohio University CTLA
• Policies for AI in Teaching and Learning from UT Austin.
• Sample syllabus AI policy statements from Penn.
• Best Practices for Transparency ffrom Barbara Komlos at the Centre for Teaching and Learning at University of British Columbia, Okanagan.
• For a one-step summary of the issues with AI detection: Heads we win, tails you lose: AI detectors in education. Bassett, M. A., Bradshaw, W., Bornsztejn, H., Hogg, A., Murdoch, K., Pearce, B., & Webber, C. (2025, October 10) or this from Michael Webb in the UK.
• Lance Eaton’s huge list of AI policies or the sortable spreadsheet.
• Sample syllabus policy statements from Northern Illinois University CITL
• Institutional Policies & Guidance collected by Joe Sabado.
• Your campus ALSO needs a clear and simple guide for everyone that lists what tools you can use and what information you can safely enter into each. Here is a great example from Tufts. Note how they clarify the types of data: public, confidential and restricted (which includes FERPA data).
• If you have any sort of institutional policy or framework for AI, you could create a customized bot for faculty on your campus to create a syllabus policy specifically for their course but aligned with your campus policy. You could write this yourself, but try this:
• Write code for an interactive interface that will produce a Generative AI syllabus policy for a college course that is both customized for individual faculty needs in that course and aligned with the university or college framework or policy. [Attach or provide a link to the campus policy or framework.] Start by asking faculty a few question about how they want to approach AI usage in their course, and offer the option of uploading a syllabus or learning goals. Offer some options based on the university framework and then provide a draft syllabus for the faculty member.

DISCLOSURE FRAMEWORKS

• Permission and Transparency in the use of Generative AI (produces a badge on use).
• The Artificial Intelligence Disclosure (AID) Framework from Kari Weaver at the University of Waterloo.
• AI Policy Code Snippet Generator from Jason Todd at Xavier University of Louisiana
• AI Transparency Statement for Assignments from Kim Acquaviva

POLICIES on AGENTIC AI

• University of Minnesota, Agentic AI Risks and User Responsibilities. For the moment, articulating risks and responsibilities seems to be the most common policy theme.
• University of Missouri System, Guidelines on AI Agents.
• Transitioning to the Agentic University 2026–27 from Ray Schroeder
• The University of Tennessee, Knoxville Agentic AI Browsers and Associated Risks
• MLA Statement on Educational Technologies and AI Agents

AI ETHICS

• Duke AI Ethics Learning Toolkit
• Cornell Ethical AI for Teaching and Learning
• Mount Holyoke College Guidelines for the Ethical Use of Generative AI
• Educause AI Ethical Guidelines
• Casey Fiesler has lots of resources and ideas including this Tech Ethics Syllabi Collection (a broad and ongoing collection of AI ethics course syllabi) and this AI Ethics & Policy News spreadsheet with tens of thousands of articles about AI by category (see the tabs at the bottom of the Google Sheet).
• ETHICAL Framework from California State University Fullerton
• Stanford CS281 (Carlos Guestrin): Ethics of Artificial Intelligence. This course site includes the syllabus, assignments, projects and reading list.
• MIT Media Lab (Blakeley H. Payne) AI Ethics Curriculum for Middle School includes hands-on activities, teacher guides, worksheets, and assessments.
• Nine Ethical Considerations from Leon Furze
• AI Ethics in Education: A Literature Review from June 2025
• “Navigating the ethical terrain of AI in education: A systematic review on framing responsible human-centered AI practices” (Dec 2024)
• UNESCO Ethics of AI
• Ethics and Governance of AI from the Berkman Klein Center at Harvard: scroll down for the major reports, academic papers, policy documents, and multimedia resources from their global dialogues series.
• The Alan Turing Institute has a workbook, AI Ethics and Governance in Practice, to support those teaching principles of AI ethics and safety in the design of algorithmic systems, with a governance framework particularly designed for public sector applications, teaching practical implementation of ethics principles
• Oxford Institute for Ethics in AI
• An early textbook is “The Ethical Algorithm” by Michael Kearns & Aaron Roth from Oxford UP.

GUIDES FOR STUDENTS

• Student Guide to Artificial Intelligence from AAC&U & Elon (English and Spanish)
• AI Agents: Guidelines and Concerns (University of Missouri System)
• Human Wisdom for the Age of AI: A Field Guide to Cultivating Essential Skills, from the AAC&U and Elon University in partnership with The Princeton Review

CAMPUS AI LITERACY PROGRAMS, MODULES & BADGES

• I created an Open-Edit List of all (90+ so far) the university microcredentials, certificates, programs and badges that Claude could find. Please correct and add yours there. So far, AI Literacy is limited to 3-4 hours of self-paced, optional online modules at most US institutions. Below are a few examples:
• University of Alabama has a free UA AI Experience: a three-hour, online micro-course that establishes a foundation of AI literacy across the UA community.
• University of South Carolina has Garnet AI Fluency 101, a 3-module self-paced course on how generative AI works, USC’s policies, and how to use AI tools like ChatGPT Edu effectively and ethically. It also provides the opportunity for an AI badge.
• Harvard now requires 3 short modules on how AI works, what it might do to your education and what it might do to society before students take Expository Writing.
• San Diego State University offers an Academic Applications of AI (AAAI) Micro-Credential  of five learning modules focusing on practical skills and takes two to four hours to complete. Participants who successfully complete the course and all activities earn a digital badge.
• Georgetown offers an AI microcredential: a self-paced, zero credit course of 8 modules that is free to students and comes with a digital badge.
• The University of South Florida offers AI Whisperer: A Microcourse in Crafting Prompts. It is a free, self-paced, 3-4 hour online course and the badge costs $39.

PROFESSIONAL DEVELOPMENT and MORE

• Dr. Aviva Legatt’s lists of AI reports, statistics and use cases are massive and kept up-to-date.
• AAC&U Institute on AI, Pedagogy, and the Curriculum online, 8-month program for campus teams to develop a project or strategy for AI with coaching, peer feedback, and monthly workshops.
• AAC&U Teaching with AI 4 Webinar Series. Four hours with José and Eddie.
• Guide to AI in Schools from MIT.
• 13 free courses about AI from MIT.
• Free self-paced Generative AI 101: Skills for Success from the University System of Georgia with micro-credential.
• Anna Mills‘s list of AI professional development opportunities, podcasts and more.
• Here is a faculty framework from Butler University.
• ACUE AI course ($179 for 1-2 hours self-paced)
• This Generative AI Product Tracker from ITHAKA lists and describes AI products marketed towards faculty and students.
• An administrative/institutional guide to AI initiatives from the University of Florida.
• A really good (new and brief) analysis of what AI might do to jobs, the economy and all of us.
• Try playing Sam Illingworth’s Bot or Not AI detection game.
  

CAMPUS STRATEGY RESOURCES

• CampusAI Exchange has lots of resources from policy to governance.
• The EAB IT Strategy Resource Hub includes 35 campus AI case studies.
• University of North Texas AI Resource Hub
• A Guide to Developing a Whole School AI Strategy from The National College for International Schools.
• Lance Eaton has created a global tracker of how higher ed accreditation agencies are changing standards and adapting to AI.
• This Gen AI Student Accusation Case Inventory from Lance Eaton offers a list of publicly-reported cases in which students have been accused of academic dishonesty using AI.

CAMPUS INITIATIVES

• Here is a huge list of Campus Use Cases from Aviva Legatt.
• Here is a list of university AI Cohort programs. 
• Creighton University Planning Guide for adding AI to Gen Ed
• The Ohio State University is integrating AI Fluency into its curriculum.
• Bowdoin College launches a $50M Hastings Initiative for AI and Humanity.
• Emerson is “formally integrating AI into the curriculum” with a new  XR-Studio in its Emerging Media Lab.
• University of Central Florida has the Institute of Artificial Intelligence, an AI for All initiative and an annual Teaching and Learning with AI Conference.
• Arizona State University has a large range of programs and tools including a CreateAI Builder tool for faculty.
• Edvisorly is one of many companies now selling admissions tools like transcript processing.
• Stanford has the Human-Centered AI Institute (HAI).
• The University of Southern California launched a $1 billion-plus initiative to expand and infuse advanced computing throughout the university’s programs and curriculum including a new School of Advanced Computing.
• Oregon State has an AI Career Chatbot trained on its own materials.