Estimated reading time: 2 minutes, 23 seconds
If you’ve been following this blog, you know I’ve written about critical AI literacy, and as I’ve keynoted, taught, and workshopped this, I’ve been developing my model further. The latest has just been published as my first LSE HE blog as a fellow there this year. (side note, I love my editor’s choice of title for the post! Go check it out!)
I’m creating this table of resources to go along with that model. I’d be very happy to learn about further resources I don’t know of for teaching each of the elements/dimensions of critical AI literacy.
| Dimension of Critical AI Literacy | Resources I recommend… |
| Understand how it works | AI Pedagogy Project AI Starter can be really useful for teachers, and LLM tutorial for learners and teachers. I also found the Practical AI for Instructors and Students playlist by Lilach and Ethan Mollick really informative. |
| Recognize inequalities and biases | Past work on inequalities and biases in AI are Joy Buolamwini’s work on racism in facial recognition AI. You can now find more of her work on a documentary on Netflix, a video of Joy in conversation with OpenAI’s CEO Sam Altman, and her recent book Unmasking AI. Others who have uncovered historical bias in AI are Safiyya Noble and her book Algorithms of Oppression, and Cathy O’Neill’s book Weapons of Math Destruction. We need to learn about “engineered inequality” in technology (Ruha Benjamin), reflecting or exacerbating so much of the inequalities in the systems within which they are designed/built (Selwyn). |
| Examine ethical issues | The European Union AI act provides a useful framing of the amount of risk associated with using AI in different fields. I really like the comprehensiveness of Leon Furze’s work on AI ethics, and I still find Autumm Caines’ work on discussing privacy policies and ethics before embarking on involving students with AI important and not sufficiently discussed. |
| Craft effective prompts | Mark Marino’s brief Secrets fo Prompting ChatGPT is a great starting point. Free prompt engineering courses such as Vanderbilt’s on CourseraA useful prompt engineering guide with multiple techniques for prompt writing is also available.Try Anna Mill’s assignment A Tale of Two Critiques to help students reflect on the differences between AI generated output and their own. |
| Assess appropriate use | Sidebar/table: 5th Dimension: Maha recommends …(see Hannigan et al’s Botshit article & Shah and Bender’s article on information access), while remaining self-aware of our own level of expertise to verify the quality and truthfulness of the output (see Aleksandr Tiulkanov, 2023 in the UNESCO Quickstart document) AI can be used maliciously to perpetuate misinformation as with deep fakes (I show students the Which Face is Real sitesee this article for more discussion on use of metaphors for critical AI literacyencouraging your students to debate whether AI use in your course/assessments is beneficial or harmful (see this idea of an activity). |