ÌÇÐÄÊÓƵ

A study in the International Journal of Mobile Learning and Organisation has mapped the fast-growing field of generative artificial intelligence (GenAI) in education. The research highlights both the technology's transformative potential and gaps in understanding how it affects learning and cognition. Academic interest in GenAI has surged, from just a handful of papers to many hundreds now.

GenAI refers to systems using deep learning and natural language processing to generate content or responses that emulate human tasks. In education, these systems are increasingly seen as capable of performing roles once the preserve of teachers, tutors, peer mentors, and administrative staff. They can deliver personalized instruction, real-time feedback, and data-driven decision support.

Applications range widely from language learning, where AI can simulate conversation and assess fluency, to health care training, where it helps students build clinical reasoning and critical thinking. Critics warn, however, of ethical issues and biases in training data that may affect fairness and trustworthiness.

The study found that most research so far centers on learner perceptions, how students feel about using GenAI, their motivation, and their trust in its usefulness. These early studies, largely survey-based, provide useful foundations for future studies but leave deeper questions unanswered. Few investigations have explored the cognitive and behavioral mechanisms underlying how learners actually interact with GenAI, a gap that could hinder efforts to design more effective and ethical systems.

Some studies have examined GenAI as a teaching support tool, focusing on human-computer interaction and adaptive learning that tailors content to individuals. Others have explored its role in assessment, offering automated feedback and grading. However, reliability remains a major concern, especially given that GenAI outputs can be inconsistent, and decision-making processes are often opaque, raising questions about fairness and academic integrity.

One term that often arises in any discussion of GenAI is "hallucinations," where the system produces what seems to be wholly fabricated output that is not based on facts.

The authors argue that more empirical research is now needed to understand how GenAI shapes learning, thinking, and evaluation. Without such evidence, there is a risk that GenAI will spread through education superficially, enhancing convenience but not necessarily deepening learning.

By mapping hundreds of studies, the paper provides valuable insights for policymakers and educators hoping to harness the potential of GenAI responsibly. It identifies both promising directions and critical blind spots in the effort to integrate AI into educational practice.