---
title: "What is AI Literacy? Competencies and Design Considerations"
authors: Duri Long, Brian Magerko
year: 2020
venue: CHI '20
kind: paper
status: note-written
captured: 2026-05-17
anchor_bucket: 3 · Academic — AI Literacy / HCI
---

# 2026-05-17 · Long & Magerko (2020) — "What is AI Literacy?"

## What it is

A widely-cited conceptual paper from CHI 2020 proposing a 16-competency framework for AI literacy. Drawn from synthesis of prior work in computational literacy, data literacy, media literacy, and existing AI-education curricula.

The 16 competencies cluster (roughly) around:

- *Know & understand AI* — what AI is, what algorithms are, AI strengths and weaknesses
- *Use & apply AI* — interacting with AI agents, prompting, programming awareness
- *Evaluate & create AI* — critical evaluation, ethical considerations
- *AI ethics* — bias, accountability, transparency

## Why it matters for our line

This is the closest academic ancestor to NOUS OS's student-facing work. When we say "AI literacy" we are walking in a frame they set.

It defines what AI-literate behaviors look like. We are downstream of it — operationalizing some of those behaviors into a measurable 20-minute loop.

## Where we share

- Same target audience (general public, with explicit K-12 attention).
- Same belief that AI literacy is multi-dimensional, not single-skill.
- Their "critical evaluation" competency directly maps to Sandbox phase 4 (source check).
- Their "AI ethics" cluster directly maps to Sandbox phase 3 (human boundary, especially privacy boundary).

## Where we differ / what we add

| Long & Magerko | NOUS OS |
|---|---|
| Descriptive 16-competency *taxonomy* | Prescriptive 6-phase *loop* with timing |
| Cognitive checklists | Behavioral protocol with measurable outcome |
| What students should know about AI | What students should do *while using* AI |
| Per-competency assessment as future work | Capability-without-AI delta as the primary outcome instrument |
| Curriculum-design oriented | Per-session-design oriented |

The key delta is **operational**: their framework lists what AI-literate behavior looks like; we propose a 20-minute behavioral protocol that produces measurable practice of those behaviors.

We do not replace their work. We sit downstream — a way to **operationalize** their competencies into a session-level instrument.

## What this changes in our practice

- We should explicitly map each Sandbox phase to a subset of their 16 competencies in `research-line.md` or a derived doc.
- Their paper gives us a defensible position for "what we mean by AI literacy" without re-litigating definitions.
- Their assessment-as-future-work suggests an explicit collaboration target: NOUS OS measurement infrastructure could be the assessment instrument their framework lacks.

## Limitations of this work (from our perspective)

- No empirical study; the paper is conceptual.
- No measurement instrument proposed for the 16 competencies.
- 5 years old (2020 — pre-ChatGPT-4); some competencies have shifted relative importance since then.
- Light on "the human's responsibility" — heavier on "the human's understanding."

## Open questions for follow-up

- Do empirical follow-ups exist that operationalize Long & Magerko's 16 competencies? (Worth a CHI / Learning Sciences search.)
- Has either author published 2024-2026 updates accounting for LLM-era shifts?
- Is there a Chinese-language equivalent / parallel framework? Worth scanning.

## Citation

Long, D., & Magerko, B. (2020). What is AI Literacy? Competencies and Design Considerations. *Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems* (CHI '20). ACM. https://doi.org/10.1145/3313831.3376727