Educators Debate AI Literacy Definitions

A community college dean's call for definitions of AI literacy drew varied responses from readers, published in a Confessions of a Community College Dean column on Inside Higher Ed. The dean noted a lack of shared understanding for the term despite widespread demand for it.

One reader submitted Ohio University's definition. It includes components of "effective practices," "ethical considerations," "subject knowledge" and "rhetorical awareness." The goal is for students to develop useful prompts, weigh data biases, understand how user expectations shape the product, and possess enough subject knowledge to judge the output.

The dean found these to be good goals but noted they stand in tension. Serious ethical objections to AI could make proficiency in using it problematic. Furthermore, students who have used AI routinely for years may lack the foundational knowledge to spot subtle biases or small hallucinations in outputs.

Proposed Metaphors and Frameworks

Another reader offered a metaphor comparing AI to a microwave. AI literacy, in this view, means understanding how to use AI as a microwave rather than the only cooking tool. The challenge is teaching this to students who may not know what a "well-made meal" tastes like, especially when many praise AI-generated results.

The dean liked this approach for making AI less threatening and recognizing its utility in specific cases. Cory Doctorow's comparison to spell-check was noted as making a similar point. The metaphor highlights anxiety about students who rely on AI for much of their education.

Formal education might provide counterexamples to ground judgments on when AI use is appropriate. Melt butter in the microwave? Sure. Cook a turkey? No.

A variation on this theme defined AI literacy as "the ability to use AI without surrendering the judgment needed to decide whether, when and how its output should be trusted." This reader suggested making it observable by having learners explain what they questioned, verified, rejected, and could still do without the tool.

This definition includes an assessment mechanism. If a student shows effective critical judgment, the technology becomes secondary. However, the dean expressed concern that effective critical judgment relies on broad context knowledge, which AI itself may threaten.

Connections to Existing Literacy Models

A thoughtful reader suggested building on the American Library Association's definition of information literacy. That definition requires individuals to "recognize when information is needed and have the ability to locate, evaluate, and use effectively the needed information." Information literacy needs skills in research and critical thinking.

The dean, a self-professed fan of libraries and librarians, found this definition too narrow for a simple substitution of "AI" for "information." AI removes much search work, but that search work is where learning often occurs. The dean suspects librarians will develop a more tailored definition soon.

Finally, Annette Vee volunteered her version of "critical AI literacy" in a separate column. She suggests it is "the ability to understand, apply and assess AI operations, uses and outputs." Her column's focus on reflection may help jump-start critical thought even with shaky subject knowledge.

The distinction between AI literacy and critical AI literacy may separate what many employers want from what educators want. Vee's work includes useful classroom exercises. The dean highly recommended reading her full column, noting it opens with a metaphor about the uses and abuses of composition courses.

The dean concluded by thanking respondents, acknowledging AI is a rapidly moving target, and stressing the obligation to students to get it right. Meaningful policy discussions require a clear sense of the goal.