Students now arrive at university having experimented with generative artificial intelligence (GenAI) tools throughout secondary school, often with wildly inconsistent guidance about what is permitted and what is not. Some have been told to avoid AI; others have been encouraged to use it freely.
For universities, this bolting horse creates a dilemma – and a mandate. Institutional policies are still evolving, disciplinary norms are uneven and faculty attitudes range from enthusiastic experimentation to cautious avoidance. In this environment, simply teaching students how to use AI tools is not enough. What students must be taught is critical AI literacy.
Critical AI literacy begins with understanding what AI is and where it comes from. Many students encounter GenAI as if the tools were neutral, when in reality they are the products of vast infrastructures of computation, data and human labour. These systems rely on enormous datasets, often assembled under ethically ambiguous conditions, and they require energy-intensive processing that carries environmental costs.
- How should universities define AI proficiency?
- GenAI practice blossoms through the open exchange of insights
- Inoculating students against AI-generated scientific misinformation
To ignore these realities is to treat AI as a kind of intellectual magic trick: impressive and mysterious. Its mechanisms should not remain unquestioned.
‘Look behind the curtain’
In response to technology’s illusions, we need to give students a robust awareness of AI’s consequences. In our research and teaching, we define critical AI literacy as the ability to:
- understand how AI works and where it came from
- appraise its strengths and weaknesses for situational and pedagogical tasks
- evaluate the ethics of its use in different situations.
Viewed this way, critical AI literacy is a way of reading the world as both data and habitat and of reminding students that every act of intelligence has an infrastructure and an accompanying cost.
GenAI excels at pattern production and synthesis. It is far less reliable at reasoning, contextual understanding or producing verifiable claims without human oversight. Students need to develop the judgement to recognise when AI can be a productive partner and when it is likely to mislead them.
This judgement is especially important in disciplines where interpretation, evidence and argument are central. In fields such as history, education or literature, AI can generate plausible-sounding narratives that mask inaccuracies. Without disciplinary knowledge and critical scrutiny, students may struggle to distinguish between convincing language and reliable insight.
The challenge is not simply technological. It is pedagogical.
Transparency and reflection
One promising approach is to emphasise transparency and reflection in student AI use. In our teaching, we allow, and even encourage, students to use AI tools, provided they disclose how they used them and reflect on how the interaction shaped their learning. One way to do this is to use a form that asks: “How did you use this tool? How did this tool influence your thinking?” for projects and assignments. Such reflection helps students recognise where their own intellectual work begins and where AI assistance enters the process. It also prevents the murky situation in which students themselves struggle to distinguish between their writing and machine-generated text – a phenomenon that is increasingly common.
In a graduate course we co-taught, for example, a student’s submitted assignment raised concerns about plagiarism. We did not rely on an AI detector; instead, our familiarity with generative tools allowed us to recognise patterns typical of AI output, including a passage that appeared to echo the original prompt. When we spoke with the student, however, the situation proved more complicated than straightforward misconduct. The student insisted the work was their own – and seemed genuinely unable to identify where their writing ended and AI-generated text began.
The case eventually went through the university’s academic integrity process and was ruled a violation, but the experience revealed a deeper challenge: if students lose awareness of the boundary between their own thinking and AI assistance, traditional ideas of authorship become harder to apply. This is why we now ask students to disclose when and how they use AI and to reflect on how it shaped their learning. It shifts the focus from policing misuse to helping students remain aware of their own intellectual contributions.
Ethical and social implications
Critical AI literacy also requires confronting the broader ethical and social implications of the technology. Generative systems raise questions about intellectual property, data extraction and environmental impact. They reshape labour markets and knowledge production in ways that students will grapple with throughout their careers.
Teaching students to engage with these issues is not about raising technological pessimism – quite the opposite. Universities should encourage thoughtful experimentation with AI, exploring how it can support creativity, feedback and self-regulated learning. But experimentation without critical awareness risks producing graduates who are highly proficient users of AI tools yet poorly equipped to evaluate them.
Higher education has always aimed to cultivate judgement as well as knowledge. AI does not change that mission; it intensifies it. Students must learn to read the world as data – to understand that every act of artificial intelligence is supported by infrastructures, assumptions and costs. The goal, ultimately, is not to produce students who simply know how to prompt an algorithm. It is to graduate individuals who can question it, challenge it and decide when it should – and should not – shape their thinking.
Amy Allen is an assistant professor of social studies the Elementary Education programme, and David Hicks is professor of history and social science education, both at Virginia Tech. Their open access book Teaching History with Chatty Geeps: A Technocurious Approach to Generative AI in the Classroom (VT Press, 2026) will be published in July.
If you would like advice and insight from academics and university staff delivered direct to your inbox each week, sign up for the Campus newsletter.
comment