Three ways to develop students’ AI literacy

Two years on from ChatGPT’s debut, the presence of generative artificial intelligence (GenAI) in higher education is impossible to ignore. It has shifted the landscape, enabling dynamic tools for learning and creativity, while exposing critical vulnerabilities in our educational frameworks. Yet, as we move deeper into the GenAI era, we must ask ourselves: are we truly preparing our institutions, educators and students to engage with this technology responsibly, or is this the optimal moment to reflect on how we can better align education with GenAI’s transformative potential?

How is higher education shaping AI literacy?

GenAI is transforming how knowledge is accessed, shared and evaluated, enabling students to draft essays, generate ideas and simulate discussions. 

• How to align AI tools with teaching philosophies: a practical guide
• AI-aware pedagogy for business courses
• Resource collection: AI and assessment in higher education

However, their ease of use risks fostering intellectual shortcuts and superficial engagement with learning. The question facing higher education is no longer whether GenAI will reshape education, but how we can integrate it without compromising foundational values such as critical thinking, academic integrity and ethical reasoning.

Early steps towards AI readiness

Early efforts to prepare us for these challenges were commendable. A 2023 free Mooc from King’s College London, followed by professional development initiatives from Jisc and the University of Cambridge, addressed foundational AI literacy, including understanding GenAI capabilities and limitations, tackling ethical concerns and integrating GenAI thoughtfully into teaching and assessment. However, while these initiatives are an encouraging start, they are largely introductory and have yet to evolve into sustained frameworks that address the full complexity of GenAI’s challenges – or at least, those we are currently aware of. 

The pedagogical strategies often advocated in institutional guidelines – such as asking students to compare AI-generated texts, analyse drafts with GenAI feedback or reflect on their learning processes – are valuable but risk falling short of fostering truly critical engagement with GenAI. These tasks often emphasise surface-level interactions, such as analysing outputs, without consistently addressing the ethical, epistemological and cognitive complexities that AI introduces. 

Without deeper interrogation of GenAI’s limitations, biases and the structural dependencies it creates, we risk normalising a technocentric view of education that prioritises functionality over critical thinking. Are we equipping students to question GenAI’s role in shaping knowledge or merely training them to work alongside it uncritically? This is a crucial distinction we must address to ensure that AI literacy evolves into AI criticality.

AI as a disruptor of assessment norms

The advent of GenAI has raised questions about the validity of traditional assessments. Essays and multiple-choice quizzes are particularly vulnerable to GenAI manipulation, rendering them less dependable as measures of student learning​. In response, some institutions have reverted to closed-book exams and other controlled conditions to mitigate GenAI’s influence. These approaches are inherently defensive and fail to engage with the broader opportunities GenAI offers for rethinking education.

A more constructive approach involves integrating GenAI directly into the learning process. At King’s, for instance, marketing students are encouraged to critically evaluate ChatGPT’s outputs while designing branding strategies. This not only develops technical proficiency but sharpens critical thinking and ethical reasoning​. Approaches like these align with the evolving needs of modern education, emphasising the application of knowledge over mere reproduction.

Consistency is key

A major challenge in integrating GenAI lies in the inconsistency of institutional policies. While some universities embrace GenAI as a teaching tool, others adopt restrictive measures that create uncertainty among students and staff​. This undermines efforts to develop coherent frameworks for AI literacy and usage.

In the UK, the Russell Group’s principles on GenAI provide a foundation for fostering GenAI literacy across UK higher education. These principles emphasise the need for universities to equip staff and students with the skills to critically engage with GenAI​. However, operationalising this vision requires more than general guidelines. Universities must invest in structured, iterative programs that go beyond introductory levels to address the nuanced challenges of GenAI integration – challenges that include fostering interdisciplinary approaches, addressing ethical dilemmas and supporting diverse learner needs​.

Teaching critical GenAI skills 

The following three strategies provide practical ways to foster these skills, ensuring that students develop not only AI literacy but the capacity for meaningful critical engagement.

1. Simulate hallucination and critique outputs. Use tools such as the Max Hallucinator to generate seemingly authoritative but flawed AI outputs. For instance, provide students with a fabricated historical analysis generated by the tool and ask them to identify errors such as non-existent events or misattributed quotes. Students should not only critique the inaccuracies but also reflect on the potential risks of trusting AI-generated content in professional or academic contexts.
2. Design ethical case studies using GenAI outputs. Create case studies where students critically assess GenAI-generated decisions with ethical implications. For example, use a GenAI tool to simulate an automated hiring recommendation that ranks candidates based on biased criteria. Ask students to identify and explain the ethical issues – such as perpetuation of systemic bias – and propose actionable solutions, such as improving the training data or implementing fairness audits. Extend the exercise by linking the discussion to relevant ethical frameworks.
3. Introduce blind spot analysis exercises. Provide students with GenAI-generated outputs that lack key perspectives, such as a summary of a global event that omits marginalised voices or environmental concerns. For example, a GenAI-produced text about the climate crisis might overemphasise industrial innovations while neglecting the disproportionate effects on vulnerable communities. Students should identify these omissions, explore why they occur (for example, biases in training data) and rewrite the outputs to include more comprehensive perspectives. This exercise not only teaches critical thinking but also highlights the importance of diverse and inclusive knowledge construction.

By incorporating these approaches, educators can move beyond surface-level GenAI literacy to foster deeper criticality in students. As higher education continues to evolve alongside AI technologies, embedding these practices into teaching will help ensure that students are not only users of GenAI but informed critics and ethical stewards of its application.

Higher education stands at a crossroads: will we empower students to think critically about GenAI’s role in shaping their future? Our response today will define how well we prepare them – and us – for the complexities ahead.

Chahna Gonsalves is senior lecturer in marketing (education) at King’s College London. Sam Illingworth is professor of creative pedagogies at Edinburgh Napier University.

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.