-
Leveling up on AI fluency
I recently went through Anthropic’s “AI Fluency: Framework & Foundations” course (free, on their Skilljar platform) and wanted to share it here since Dr. Ron thought it’d be a good use of our Groups feature to start these kinds of discussions.
The course centers on a “4D Framework” for working with AI tools like Claude, ChatGPT, etc, Delegation (deciding what to hand off to AI vs. do yourself), Description (how you prompt/frame a task), Discernment (evaluating what the AI gives back instead of trusting it blindly), and Diligence (the habits you build around checking, verifying, and using AI responsibly over time).
A few things that stuck with me: AI models are fundamentally predicting likely text, not looking facts up, which is exactly why hallucinations concentrate in specific details like names, dates, and citations rather than general explanations. Also useful: models have a fixed knowledge cutoff, and their “working memory” doesn’t degrade gradually, but it holds up fine until it hits a hard limit, then performance drops off a cliff.
Given how much AI is coming up in our community internet and digital equity work, I think this is worth a group conversation.
