ChatGPT in the Lecture Hall: Nigerian Students Weigh AI’s Promise Against the Price of Thinking for Themselves

Generative artificial intelligence has swept into university classrooms around the world faster than most institutions can write policies about it, and nowhere is the tension more visible than in Nigeria’s public universities. A new qualitative study published in Discover Education offers one of the most detailed looks yet at how undergraduates in the country are living with these tools, interviewing twelve computer science students drawn from six public universities across all of Nigeria’s geopolitical zones. The findings paint a picture that is neither utopian nor dystopian, but something more interesting: a generation of students actively negotiating, on a daily basis, whether AI is making them smarter or quietly doing their thinking for them.

The research, conducted by Hammed Adedeji Ajani and Sam Ramaila of the University of Johannesburg, is grounded in two theoretical frameworks that give the work unusual analytical depth. The first is 21st Century Competency Theory, which holds that learners need more than content knowledge; they need the capacity to apply what they know ethically, creatively and collaboratively in fast-changing environments. The second is Vygotsky’s Sociocultural Learning Theory, which treats learning as a socially mediated process shaped by cultural tools. Within that lens, generative AI becomes something more than software: it is a cultural tool that can operate inside a learner’s zone of proximal development, enabling students to complete tasks they could not manage alone, but only if the scaffolding is used rather than substituted for genuine cognitive work.

Methodologically, the study took an interpretive phenomenological approach, combining in-depth semi-structured interviews lasting 25 to 38 minutes with a 95-minute focus group discussion. The twelve participants, seven men and five women, were selected through a multistage sampling design: the researchers first chose one university from each geopolitical zone by systematic random sampling, then purposively invited two leaders of each institution’s Computer Science Students’ Association who had extensive experience with digital tools. Thematic content analysis organised the data into three overarching themes covering exposure and adoption, the enhancement or hindrance of competencies, and the broader perceived impact of AI on future skills. The study received ethical clearance from the University of Ilorin’s review committee, and member checking, an audit trail and reflexive journaling were used to strengthen credibility.

The first striking finding concerns how students encounter AI in the first place. Awareness rarely came from official channels. Instead, participants traced their first exposure to social media platforms such as Twitter, TikTok and YouTube, where influencers and technology enthusiasts demonstrated tools like ChatGPT; to lecturers introducing the technology during classes on emerging technologies; to peer recommendations during coursework; and to explorations of coding platforms like GitHub. One student was introduced to ChatGPT at a tech workshop organised by a university computer science department. In other words, the pipeline into AI adoption in Nigerian higher education currently runs through informal social networks at least as much as through formal instruction, a pattern that maps neatly onto the sociocultural framing of the study.

Once adopted, the tools quickly became embedded in academic routines. ChatGPT emerged as the dominant platform, prized for its accessibility and versatility, with Gemini, Microsoft Copilot, Grammarly and GitHub Copilot also in regular use, and some students experimenting with image generators such as DALL-E and MidJourney. Roughly half of the participants reported using AI tools multiple times a week, particularly for essay writing, research summarisation and code generation, while a few admitted near-daily use during exam periods and intensive projects. The applications were strikingly varied: summarising complex research papers, debugging programs, generating sample scripts, explaining difficult topics in data structures and algorithms, brainstorming research ideas, conducting literature reviews, formatting references, translating languages, building presentations and even generating study schedules for time management.

Crucially, the students reported that this engagement was cultivating genuinely future-oriented skills. Several said interaction with AI had introduced them to prompt engineering, the emerging craft of designing effective instructions for language models. Others credited AI-assisted programming with improving their coding proficiency, data visualisation techniques and automation skills, while one participant described refining professional reports with AI and thereby strengthening workplace communication; another said the tools offered insights into user interface and user experience design principles. These are precisely the cognitive, technical and interpersonal capabilities that employers increasingly demand as industries absorb artificial intelligence, and they align closely with the competency framework underpinning the research.

Yet the same interviews surfaced a catalogue of concerns that any institution hoping to integrate AI responsibly must confront. Accuracy and reliability topped the list: students reported that AI outputs can be outdated, incorrect or vaguely misleading, particularly in technical subjects where the models lack contextual understanding. One participant noted that AI-generated citations are often fabricated, forcing manual cross-checking of every reference before use. Ethical anxieties were equally prominent, with some lecturers discouraging AI use altogether, leaving students uncertain about where legitimate assistance ends and academic dishonesty begins. And overlaying everything was the issue of equity: poor internet connectivity in areas with weak network coverage, and the cost of accessing advanced tools, mean that the benefits of AI are distributed unevenly across the student population, deepening an existing digital divide.

The most philosophically charged findings concern the effect of AI on thinking itself. Most participants acknowledged that generative AI had sharpened their critical thinking and problem-solving, helping them break down complex problems, refine research questions and consider alternative perspectives. But others worried aloud that instant answers invite shallow engagement. One student observed that the technology sometimes gives quick answers that reduce personal reflection on problems; another argued that instant AI-generated solutions might erode the need for deep reflection, human interaction and original thought. On creativity, opinions split: some found that AI suggested ideas and frameworks they would never have conceived alone, while others feared that over-reliance on machine-generated suggestions could stifle originality. Several students described deliberate strategies to preserve their independence, such as using AI purely as a brainstorming partner while ensuring final ideas were shaped by personal input, or consciously limiting use to avoid over-dependence.

Career preparedness added a further layer of ambivalence. Many students believed AI equips them with relevant digital skills and keeps them current with industry trends, accelerating prototyping in programming and content creation. But sceptics cautioned that AI’s rapid evolution could render current knowledge obsolete if foundational skills are not solid, and that enhanced productivity might come at the cost of deep expertise in certain fields. Ethical considerations, including plagiarism, misinformation and the tendency of AI systems to reinforce biases present in their training data, were widely acknowledged, with several students insisting that AI responses must always be verified against reliable sources before being accepted as fact.

The authors’ conclusion is that generative AI’s impact on future competencies is not inherently positive or negative; it depends on how the technology is adopted, regulated and integrated into learning. They call for structured, policy-guided curricular integration that embeds AI literacy, ethical awareness and critical evaluation skills, alongside clear institutional policies that safeguard academic integrity, authentic learning opportunities such as research projects and innovation challenges, and concrete efforts to improve access to AI tools and digital infrastructure for students facing financial or connectivity barriers. The study’s limitations are acknowledged candidly: a small, purposively selected sample of technically oriented students, reliance on self-reported data, and no capacity to capture long-term effects. Even so, the message from Nigeria’s lecture halls is clear. Students are not passive consumers of artificial intelligence; they are its most perceptive critics, and they are asking the same question their universities must now answer: how to harness a powerful cognitive tool without surrendering the very skills that make thinking worthwhile.

Subject of Research: Generative AI adoption and its impact on 21st-century competency development among Nigerian public university undergraduates

Article Title: Generative AI vs. future competencies of Nigerian public university students

Article References: Ajani, H. A., & Ramaila, S. (2026). Generative AI vs. future competencies of Nigerian public university students. Discover Education, 5(1), Article 1011. https://doi.org/10.1007/s44217-026-02218-x

Image Credits: AI Generated

DOI: 10.1007/s44217-026-02218-x

Keywords: generative AI, higher education, Nigeria, 21st century skills, critical thinking, ChatGPT, academic integrity, digital divide, sociocultural learning theory, computer science students, curriculum integration, prompt engineering