Why AI Literacy in Schools is the New Critical Thinking

The results are in, and they are telling.. The latest EU Kids Online (EUKO) 2025 report, which surveyed 25,000 children across 20 countries, including Croatia, confirms what teachers see every day: AI is no longer a “future” concept. It is already embedded in the lives of 9-to-16-year-olds, reshaping how they write homework, search, create, and communicate. AI literacy in schools shouldn’t be luxury. While students are already using generative AI daily, there is a massive difference between operating a tool and understanding its output.

True AI literacy isn’t about learning “prompt engineering.” It is about digital citizenship and AI, understanding how these models think, where they fail, and why they can be confidently wrong. We spend plenty of time talking about how students use AI, but almost no time teaching AI literacy through a critical lens.

It shows in every classroom. Students paste prompts into ChatGPT, receive a polished response, and move on. They don’t verify sources (which are often “hallucinations”), they don’t question the logic, and they don’t check for contradictions. They accept it because it sounds right.

This is a critical thinking problem that AI has simply made impossible to ignore. To foster responsible AI use in education, we must stop treating AI as a search engine and start treating it as a persuasive, but fallible, co-pilot.

Back in 2015, Robert Ennis published a framework outlining 15 skills every critical thinker needs. These include:

• Analyzing arguments.
• Judging the credibility of sources.
• Spotting unstated assumptions.
• Dealing with logical fallacies.
• Thinking suppositionally.

These 15 skills must be applied directly to AI use in the classroom. Every time a student engages with Claude, Gemini, ChatGPT, or any other AI tool, they should be mapping these skills to the output. We aren’t just teaching them to write a prompt; we are teaching them not to surrender their thinking in the process.

To move beyond passive consumption, we can apply Robert Ennis’s 2015 critical thinking framework. These 15 skills map perfectly to evaluating AI outputs:

Logic, assumptions, and contradictions

Students must learn to spot unstated assumptions in an AI’s reasoning and handle “equivocation”, where the AI uses a word in two different senses to bypass a logical flaw.

Credibility and source checking

AI literacy requires judging the credibility of sources. If an AI provides a citation, the first step of critical thinking and AI is verifying that the source actually exists and supports the claim.

Bias, framing, and missing context

Every model has a “worldview” based on its training data. Students need to ask: What is missing? What bias is inherent in this framing?