The Problem Didn't Start With AI

The workload crisis predates any conversation about artificial intelligence.

Over the past decade, international education systems have raised expectations around planning, assessment, differentiated learning, and curriculum alignment. In Ireland, the shift towards more teacher agency and student-centred learning has sound pedagogical goals. But they have been introduced without providing the tools or time to deliver them.

Teachers have absorbed the gap. Planning that should happen during the working day gets pushed into evenings and weekends. Administrative tasks expand to fill time that does not exist. And the work that drew most people into teaching, like engaging with students, is increasingly squeezed by everything that surrounds it.

This is the structural reality into which all new technology arrives. It explains why teachers' response to AI is simultaneously hopeful and wary. The promise is that it could ease this burden. The fear is that it will add another layer of complexity to an already stretched profession.

The Distinction

In recent weeks, the conversation has sharpened. A prominent piece in the Irish Times argued that AI is a threat to the essence of education. The OECD's Digital Education Outlook report puts it plainly: an overreliance on AI risks "turning students into passive consumers and teachers into supervisors." Teachers' unions have raised pointed questions about how AI-generated work will be authenticated under the new Leaving Certificate Additional Assessment Components, with 88% of surveyed ASTI members saying these assessments will create difficulties in this area.

But much of the debate treats AI as a single, monolithic category, as though a teacher asking ChatGPT for lesson ideas and a student using it to write an essay are the same thing. They are not, and the truth is, this debate is much deeper. Until we start making these distinctions, the conversation will remain a mile wide and an inch deep. This blog intends to do that. We will take these issues over the coming months and explore each in depth.

The Problem with Generic AI in Education

General-purpose AI tools like ChatGPT are impressive. They can generate lesson plans, suggest assessment ideas, draft student resources, and answer pedagogical questions with remarkable fluency. Teachers are resourceful, and many have found creative ways to put these tools to work.

But generic tools carry inherent limitations in an educational context. They don't seem to understand the structure of the Irish curriculum. They cannot reliably map a learning outcome to the relevant specification. They do not know how a Junior Cycle assessment standard connects to a particular strand or element.

When teachers use generic tools, they take on additional cognitive work to check, adapt, and align the output. In many cases, this adds time rather than saving it, which is precisely the concern the ASTI's survey identified.

It also connects to the issue of teacher autonomy and agency. When a teacher uses a generic tool, they become responsible for quality-assuring output that may look authoritative but may not be pedagogically sound. The teacher is not supported by the technology. They are supervising it. This is a meaningful difference and one that echoes the OECD's warning about teachers becoming “supervisors” rather than what we should be developing: teachers as professional agents who autonomously work with AI.