Beyond knowing how to use the tools
The University of Leicester is making the case that AI literacy belongs in the core curriculum, and crucially that such literacy means far more than knowing how to operate the tools. The university frames AI literacy as a critical, ethical, and context sensitive capability rather than simply familiarity with AI systems. This is an important distinction, and it cuts against a common assumption. As generative AI becomes ubiquitous in education, there is a temptation to treat AI literacy as a matter of teaching students which buttons to press, how to write effective prompts, and how to get useful output. Leicester argues this is precisely the wrong frame.
Professor Xue Zhou, who co authored the underlying research, put the principle directly. AI literacy should be understood as critical, ethical, and context sensitive capability rather than simply familiarity with AI tools, she said. The argument is that operating the tools is the easy and superficial part. The hard and important part is developing the judgment to use them well, to know when to trust their output and when to question it, to understand their limitations, and to wield them ethically. That judgment does not emerge automatically from use, which is the central and somewhat uncomfortable finding the work advances.
What the research found
The recommendation rests on empirical work, not assertion. A research paper titled The agency gap, examining perceived human AI agency, reflection, and generative AI learning across UK and China based higher education contexts, was published in Studies in Higher Education on June 11. The study analyzed 309 university students, 145 based in the UK and 164 in China, surveyed between September and December 2025. The cross cultural design, spanning two very different educational systems, strengthens the findings by suggesting the patterns are not artifacts of a single national context but reflect something more general about how students learn with AI.
The core finding is that interacting with AI is not enough on its own to build critical thinking. As Zhou explained, this suggests that simply interacting with AI is not enough, and that students need opportunities to reflect. The distinction between using a tool and reflecting on that use is the heart of the matter. A student can use AI to complete assignments, get answers, and produce work without ever developing deeper understanding or sharper judgment, potentially even weakening those capacities through over reliance. The research indicates that structured reflection on how and why AI is being used is what converts tool use into genuine learning, and that without it, use can substitute for thinking rather than supporting it.
The danger of cognitive offloading
The implicit warning in this research deserves to be stated plainly. There is a real risk that AI tools, used without reflection, allow students to offload their thinking rather than develop it. When a student can get a well written essay, a worked solution, or a polished analysis from an AI system instantly, the cognitive effort that once produced learning can be bypassed entirely. The student gets the output without doing the thinking that the output was supposed to represent and, more importantly, without the mental work through which understanding is actually built. The product appears, but the learning that the product was meant to evidence does not.
This is among the most serious concerns in education's encounter with AI, and it is easy to underestimate. The convenience of AI assistance can quietly undermine the very learning that education exists to produce, and it can do so invisibly, because the work looks done. Leicester's emphasis on reflection is a direct response. By requiring students to reflect on their use of AI, to consider what the tool did, what they contributed, and what they actually understand, educators can ensure that the technology supports learning rather than replacing it. Reflection forces the cognitive engagement that unreflective use allows students to skip, and that engagement is where understanding lives.
Embedding it in the core, not the margins
Leicester's call to embed AI literacy in core curricula, rather than treating it as an optional add on or a one off workshop, reflects a clear and defensible judgment about how important this capability has become. Relegating AI literacy to the margins, an elective module here, a guest lecture there, signals that it is peripheral, a nice to have rather than essential. Embedding it in the core curriculum signals the opposite, that the ability to work critically and ethically with AI is now a fundamental capability every graduate needs, comparable to writing or quantitative reasoning, regardless of their field of study.
This is the correct framing for a world in which AI will be woven into nearly every profession. Graduates will encounter these tools in whatever work they go on to do, and their ability to use them with judgment rather than naive trust will shape their effectiveness and their professional integrity. Treating AI literacy as core acknowledges that reality. The harder question, which Leicester is beginning to confront, is how to do it well across an entire institution, given that embedding a genuinely critical and reflective capability into every program is far more demanding than adding a module about which tools exist and how to prompt them.
Backed by real engagement
Leicester's recommendations are grounded in sustained institutional activity rather than abstract enthusiasm, which lends them credibility. The university reported substantial engagement with its AI education efforts, including an AI and Robotics Symposium that drew more than 300 registrations and around 200 participants, Teach with Tech sessions with 182 attendees across nine sessions between February 2025 and March 2026, an AI and higher education lecture series that attracted 474 attendees across seven online lectures, and an AI Community of Practice with 240 members. These figures suggest a real and ongoing commitment, not a single initiative dressed up as a strategy.
This grounding matters because it distinguishes thoughtful institutional engagement from reactive box ticking. Many universities are scrambling to respond to AI, often with hastily assembled policies and superficial guidance produced under pressure. Leicester's combination of original research and sustained programming suggests a more considered approach, one that takes seriously the difficult questions about how AI changes learning rather than simply reacting to the immediate disruption of students using chatbots to write essays. The depth of the engagement is itself part of the argument, demonstrating that the institution has thought about the problem long and hard enough to have something substantive to say.
A model for higher education
Leicester's approach offers a model that other institutions would do well to study, because the questions it confronts are universal across higher education. Every university is wrestling with how to respond to generative AI, and the responses so far have ranged from outright prohibition to uncritical embrace, neither of which serves students well. The reflective, critical, embedded approach that Leicester advocates charts a more thoughtful middle path, one that neither pretends AI can be banned from education nor surrenders to the convenience that lets students bypass learning altogether.
The broader lesson extends well beyond universities to anyone thinking about AI and human capability. The research finding that tool use alone does not build judgment, and that reflection is what converts use into genuine understanding, applies to workplaces and professional development as much as to classrooms. As AI becomes ubiquitous, the organizations and individuals that thrive will be those who use it reflectively, maintaining and sharpening their own critical capabilities rather than allowing the tools to erode them. Leicester's contribution is to provide evidence for what many have sensed intuitively, that the goal is not just learning to use AI, but learning to think alongside it without letting it think in your place.
