Sara Mashayekh, School of Education, UNSW Sydney
For the past few years, much of the conversation about generative AI in education has focused on adoption. Which tools should we use? How should students use them? What policies do we need? What does responsible AI use look like?
These were necessary questions. Much of the work on AI literacy has also understandably focused on developing people’s capacity to understand, use and critically evaluate AI.
But AI is changing, and the educational conversation needs to change with it.
AI is no longer simply finished software that educators either adopt or reject. Increasingly, we can modify instructions, establish boundaries, customise systems and influence how AI responds in particular contexts. That creates a different possibility.
Arguably, the next frontier is not AI adoption, it is AI design.
Moving beyond the tool
A common starting point for educators is: Which AI tool should I use?
That made sense when our main task was learning what these technologies could do. But as AI becomes increasingly configurable, simply selecting a tool tells us less and less about the educational experience that will follow.
The same underlying technology can be used in very different ways. In one learning experience, AI might largely provide information. In another, its use might encourage learners to question, compare, revise or reflect.
The technology may be identical. The educational experience is not.
This is consistent with a much longer tradition of understanding teaching as a design activity. Technologies do not determine educational value by themselves. What matters is what educators create around them.
From prompting to designing
Prompting will remain useful. There is obvious value in knowing how to communicate effectively with generative AI. But prompting and designing are not quite the same thing.
Prompting tends to ask: How can I get a better response from this system?
Design raises a broader question: How can this technology be shaped in ways that make sense educationally?
That distinction becomes more important as the systems themselves become easier to customise. Educators increasingly have options beyond accepting a generic chatbot exactly as it is.
This also means that better AI does not automatically mean better education. More capable systems create more possibilities, but educational value still depends on the choices made around their use.
Research on AI assistance and student agency, for example, reminds us that the way AI is incorporated into an activity can affect what learners themselves continue to do. The important issue is therefore not simply whether AI is available, but what its use changes about the educational experience.
Adapting AI to education
This shift has implications for educators.
Until now, much of our attention has gone into adapting educational practice to AI. We have learnt new tools, reconsidered assessment, developed guidance and policies, and tried to understand what responsible use looks like.
Those responses remain necessary. But we are reaching a point where the relationship can begin to move in the other direction.
Rather than continually asking education to accommodate the capabilities and conventions of new AI systems, educators can increasingly influence how those systems are configured for educational contexts.
UNESCO’s AI competency framework for teachers similarly points towards a more active role for educators in critically and creatively shaping AI enabled educational practice.
This is more than a technical shift. It suggests a different relationship between educators and technology. Educators are not simply users of educational AI. Increasingly, they may become its designers.
A new conversation
AI literacy will remain essential. Educators need to understand AI, recognise its limitations, evaluate its outputs and consider its ethical implications.
But the growing configurability of AI raises another question for the sector. What capabilities will educators need when they are no longer simply choosing and using AI systems, but increasingly shaping them for educational contexts?
We are only beginning to answer that question.
For the first phase of generative AI in education, educators were largely asked to adapt to AI.
Perhaps the next phase should reverse the relationship.
How should we adapt AI to education?
That may be where the most interesting educational work begins.
References
Darvishi, A., Khosravi, H., Sadiq, S., Gašević, D., & Siemens, G. (2024). Impact of AI assistance on student agency. Computers & Education, 210, 104967. https://doi.org/10.1016/j.compedu.2023.104967
Laurillard, D. (2012). Teaching as a design science: Building pedagogical patterns for learning and technology. Routledge. https://doi.org/10.4324/9780203125083
Miao, F., & Cukurova, M. (2024). AI competency framework for teachers. UNESCO. https://doi.org/10.54675/ZJTE2084
Ng, D. T. K., Leung, J. K. L., Chu, S. K. W., & Qiao, M. S. (2021). Conceptualizing AI literacy: An exploratory review. Computers and Education: Artificial Intelligence, 2, 100041. https://doi.org/10.1016/j.caeai.2021.100041
Bio
Sara Mashayekh is a Lecturer in the School of Education at UNSW Sydney. Her research examines how educational environments can be designed to develop human judgement, agency and meaningful participation, with particular interests in human centred AI, educational design, researcher and teacher development, and inclusive education.