Add TechGraph on Google
Generative AI is rapidly entering classrooms, boardrooms, and training programs. Yet a critical question remains unresolved: does AI enhance learning, or quietly replace it?
In my recent research published in the International Journal of Management Education, I examine a faculty-led model of integrating generative AI into executive economics teaching. Conducted across eight Executive MBA cohorts, the study finds that structured use of AI, when guided by faculty, significantly improves student performance, particularly among lower-performing learners.
This is not a story about AI replacing educators. It is a story about what happens when AI is carefully designed into the learning process.
The Problem with “Plug-and-Play” AI in Education
Much of the current conversation around AI in education is driven by tools: AI tutors, automated grading, or content generation systems. These solutions promise efficiency but often miss the core purpose of education: developing judgment, not just answers.
In executive education, this challenge is even sharper. Participants are not passive learners; they are decision-makers. They bring experience, context, and real-world constraints. A generic AI tool, however powerful, cannot replicate that nuance.
The risk is subtle but important: when AI becomes an answer engine, learners may outsource thinking rather than deepen it.
A Different Approach: AI as a Thinking Partner
The model I tested takes a different route. Instead of treating AI as a shortcut, it positions AI as a Socratic partner, a tool that supports questioning, refinement, and reflection.
In practice, this involved three key design elements.
First, structured prompting: students interacted with AI through guided, scenario-based questions tied to real business dilemmas.
Second, iterative dialogue: AI was used not to generate final answers, but to test reasoning, challenge assumptions, and refine arguments.
Third, annotated transcripts: students submitted their AI interactions, explaining where they agreed, disagreed, or revised their thinking.
This last step is critical. It shifts AI use from passive consumption to active engagement. Students are not just using AI; they are learning how to think with it.
What the Data Shows
The results were clear.
Students in AI-supported cohorts scored, on average, 2.45 points higher (out of 40) than those in traditional cohorts. More importantly, the gains were not evenly distributed.
Lower-performing students benefited the most.
Analysis across different performance levels showed that AI support had the strongest impact at the lower end of the distribution, suggesting that structured AI use can act as an equaliser in learning outcomes.
This is a significant insight. Much of the fear around AI in education focuses on unfair advantage or misuse. But when designed properly, AI can reduce disparities rather than widen them.
Why Faculty Still Matter, More Than Ever
One of the most important findings from the study is that the impact did not come solely from AI. It came from how AI was integrated.
The same tool can produce very different outcomes depending on instructional design. Without structure, AI leads to surface-level summaries. Structure enables deeper reasoning. This places faculty at the centre of the process. Their role evolves, but it does not diminish. Instead, it expands.
Faculty become prompt designers, framing the right questions that guide learning. They act as cognitive coaches, helping students interpret and challenge AI outputs. And they serve as ethical anchors, ensuring transparency, authorship, and responsible use. In this sense, the “human-in-the-loop” is not a limitation; it is the source of value.
Implications Beyond the Classroom
While this study is grounded in executive economics education, the implications extend much further.
In corporate training, leadership development, and workplace decision-making, AI is increasingly used as a support tool. The same principle applies: AI creates value not when it replaces thinking, but when it structures better thinking.
For organisations, this suggests a shift away from simply adopting AI tools toward designing AI-enabled processes. The focus should be on how individuals interact with AI, not just whether they use it.
For educators and institutions, it highlights the need for intentional integration rather than reactive adoption driven by technological trends.
The Hidden Layer: Behavioural and Cognitive Impact
There is also a behavioural dimension that deserves attention.
Students initially expressed hesitation, uncertainty about ethics, over-reliance, or even the legitimacy of using AI in learning. However, when guided properly, this hesitation evolved into confidence and curiosity.
This aligns with behavioural economics. Tools alone do not change outcomes; structure, framing, and incentives do. In this context, AI acts as a nudge. It can either push learners toward deeper engagement or reinforce shortcuts, depending on how it is designed.
A More Realistic Future for AI in Education
The future of AI in education will sit between full automation and traditional teaching. The question is not whether AI should be used, but how.
This research points to a simple direction: keep AI in the loop but faculty in control; design for thinking, not just output; and focus on learning, not just engagement. In an algorithm-driven world, the ability to question, interpret, and apply knowledge becomes more valuable, not less.
But beyond performance lies responsibility.
A sustainability mindset in AI-enabled learning calls for three shifts. Cognitively, learners must question AI outputs rather than accept them at face value. Behaviourally, they must use AI to deepen reasoning, not replace it. Affectively, they must remain aware of how their decisions shape organisations, society, and long-term outcomes.
When these come together, AI moves from a productivity tool to a decision-making partner grounded in responsibility. The real opportunity is not just to teach students how to use AI, but to shape how they think, act, and decide with it.
That is where human judgment remains essential.