Using AI is not the same as understanding it. The gap between the two is where the problems are.
Literal meaning: AI literacy is the capacity to understand how AI systems function — their training processes, limitations, failure modes, and structural biases — in addition to being able to use them as tools. It is distinct from AI proficiency: you can be proficient without being literate.
Origin: The concept developed in educational and policy discourse from the mid-2010s, gaining urgency with the mainstreaming of generative AI from 2022. Researchers including Meredith Broussard (Artificial Unintelligence, 2018) and Kate Crawford (Atlas of AI, 2021) established the analytical foundation: AI systems embed assumptions, reproduce biases, and have structural limitations that users who only interact with outputs cannot see. UNESCO published a global framework for AI literacy in 2022.
The capacity to understand how AI systems work, fail, and are shaped by their training — not just to use them.
The Appeal: AI literacy is framed as both civic necessity and professional advantage. In education, it offers a way to engage with AI tools critically rather than passively. For policymakers, it underpins informed regulation. The concept is widely supported because it is genuinely useful.
The Friction: The term risks being emptied of content. “AI literacy” in many educational and corporate contexts means using AI tools confidently — which is proficiency, not literacy. AI Hallucination — fluently false output — and Sycophancy (AI) — structured agreement with user premises — are exactly the failure modes that literacy should address. But if literacy reduces to tool use, these remain invisible. AI Dependency — structural reliance without critical reflection — is what low-literacy AI adoption produces at scale.
Why This Matters: AI literacy is the term that most directly determines whether the other terms in this cluster are legible. Without it, hallucination is just an error, sycophancy is just a feature, and dependency is just efficiency.
Related terms: AI Hallucination · Sycophancy (AI) · AI Dependency · Cognitive Offloading · Deskilling · AI-Ninja · Prompt Engineer
Read more:
- Artificial Unintelligence — Broussard, M. (2018). MIT Press
- What is AI Literacy? Competencies and Design Considerations — Long & Magerko (2020), ACM CHI