Artificial Intelligence | ChatGPT for Instructors

Bias

• AI Expert Meredith Broussard: ‘Racism, Sexism and Ableism are Systemic Problems’ . The Guardian.  Racism, sexism and ableism are systemic problems that are baked into our technological systems because they’re baked into society. It would be great if the fix were more data. But more data won’t fix our technological systems if the underlying problem is society.
• Language Models Might be Able to Self-Correct Biases—If You Ask Them.  MIT Technology Review. A study from AI lab Anthropic shows how simple natural-language instructions can steer large language models to produce less toxic content.

• A Radical Plan to Make AI Good, Not Evil. OpenAI competitor Anthropic, on using AI to train AI (instead of humans). They call it “constitutional AI.”

Copyright 

• Generative AI Meets Copyright. Pamela Samuelson, Professor of Law, UC Berkeley (YouTube). Recording of a talk she gave, summarizing the outstanding cases. "It will take years for this to be resolved in the courts."
• Prof. Matthew Sag Testimony on Copyright and AI (PDF). Testimony before the U.S. Senate Committee on the Judiciary, Subcommittee on Intellectual Property, July 12, 2023.
• Letter to the U.S. Copyright Office from the Library Copyright Alliance. The American Library Association and The Association of Research Libraries. Oct, 31, 2023. Download the PDF.  Supports the idea that training data for generative AI should be considered fair use.

Identifying AI Created Content

• Content Credentials Verification Tool. Upload an image, find out whether is has metadata from AI image generation tools from OpenAI, Microsoft, and Adobe. Not foolproof because other image generators don't include this metadata.

• Not by AI.  Ideas for a badge for human created content. 
• Watermarking ChatGPT, DALL-E and Other Generative AIs Could Help Protect Against Fraud and Misinformation. Berkeley iSchoolGenerative AI is forcing people to rethink what it means to be authentic. The Conversation.

Postplagiarism

• Eaton, S.E. Postplagiarism: Transdisciplinary Ethics and Integrity in the Age of Artificial Intelligence and Neurotechnology. Int J Educ Integr 19, 23 (2023). https://doi.org/10.1007/s40979-023-00144-1

• Future of Writing in the Disciplines and Professions. White paper from Carnegie Mellon.

• Four Singularities for Research: The Rse of AI is Creating Both Crisis and Opportunity. Ethan Mollick

• How the rise of AI is affecting Writing, Publishing, and Research.  AI & Accessibility. Center for Teaching Innovation, Cornell University.

• Using AI to Help Students Teach in Order to Learn. Inside Higher Ed. "By changing ChatGPT’s system prompt, we can create content misunderstandings that students can correct, write Joel Nishimura and Anna Cunningham."

• New BYU Computer Science Study Shows Four Ways Students are Actually Using ChatGPT. Brigham Young University News. A survey finds 4 categories of use.

• Teaching and Generative AI: Pedagogical Possibilities and Productive Tensions. Beth Buyserie  & Travis N. Thurston.  Open access ebook on Pressbooks.

• Instructors as Innovators: a Future-focused Approach to New AI Learning Opportunities, With Prompts. Ethan Mollick & Lilach Mollick. A paper that explores how instructors can leverage generative AI to create personalized learning experiences for students that transform teaching and learning. 

• AI Detection in Education is a Dead End. Leon Furze. How AI detection tools work and why they don't work.

• Generative AI in Higher Education: The ChatGPT Effect. Cecilia Ka Yuk Chan, Tom Colloton. Open access ebook. Explores the strengths and weaknesses of incorporating ChatGPT into curriculum design.

• Incorporating Generative AI in Teaching and Learning: Faculty Examples Across Disciplines. Columbia University Center for Teaching and Learning. Faculty across disciplines provide a glimpse into their approaches as they experiment with AI in their classrooms.

• Handout: AI and the Future of Teaching and Learning. U.S. Dept of Education, Office of Educational Technology (PDF). Two-page list of key insights and recommendations from the U.S. Dept. of Education.

• Understanding AI Writing Tools and their Uses for Teaching and Learning at UC Berkeley.  Berkeley Center for Teaching and Learning. 

• Critical AI, New Interdisciplinary Journal. Rutgers and Duke. Follow the new journal at the link provided, or check out the blog feed site including, “research, reviews, and commentary by interdisciplinary scholars in a wide range of AI-adjacent fields, as well as posts by faculty and students affiliated with the Critical AI @ Rutgers initiative.”

• Special issue: Critical AI a Field in Formation, American Literature.  Duke “This special issue provides an overview of the emerging interdisciplinary field of Critical AI, which seeks to demystify artificial intelligence; counter its mythologizing as a marvelous and impenetrable black box; and translate, interpret, and critique its operations, from data collection and model architecture to decision making. Artists and researchers are developing new methods, practices, and concepts for this critical project, which is both historicist and attentive to the institutional, technological, and epistemic transformations still underway.”

• Elements of AI. University of Helsinki. Free course for learning about AI in general. 2024.

• ChatGPT is Everywhere. Here’s Where it Came From.  Will Douglas Heaven, MIT Technology Review.  Provides a useful history.

• The Acceleration of Artificial Intelligence.  Anna-Sofia Lesiv.  An interesting summary of the development of today’s generative AI. Useful for understanding the 2017 breakthrough (transformer architecture) that led to so much that’s happening now.

• A Jargon-Free Explanation of How AI Large Language Models Work. Timothy B. Lee and Sean Trott. Learn about word vectors, transformers, and more.

• Generative AI Exists Because of the Transformer. Visual Storytelling Team and Madhumita Murgia, Financial Times. Useful explanation with helpful visuals.

• Why Does ChatGPT Make up Fake Academic Papers? Twitter thread by David Smerdon, Univ. of Queensland. Very good explanation.

• Large Language Models from Scratch and Large Language Models: Part 2 (videos). Graphics in 5 Minutes on YouTube.

• What is Reinforcement Learning From Human Feedback (RLHF). Alex McFarland. Useful explanation.

Policies and FAQs from OpenAI (worthwhile reading):

• How Your Data is Used to Improve Model Performance
• Privacy Policy | API Data Privacy
• How ChatGPT and Our Language Models Are Developed

• Prompt Engineering for ChatGPT. Coursera course by Dr. Jules White, Vanderbilt University.

• How to Use AI to Do Practical Stuff: A New Guide. Prof. Ethan Mollick.

• Three Ways to Leverage ChatGPT and Other Generative AI in Research. Times Higher Education. Ideas for using ChatGPT for (1) to determine a hypothesis or question; (2) research method: follow an accepted research method or invent a new method or algorithm to conduct an investigation that resolves or answers the question; and (3) research output: formulate, evaluate, and document the solution to enable further research.