Could AI end grind schools?

AI in education is not the Messiah, but it could improve things, especially where resources are tight 

Jane D'Alton (Head of Pastoral and Safeguarding) at the Institute of Education on Leeson Street in Dublin welcomes students for the beginning of the Leaving Certificate exams last month. International school group Dukes Education purchased the Institute in 2023 for a sum close to €130m. Photo: Justin Farrelly / ©RollingNews.ie

With ASTI’s rejection of proposals for Leaving Certificate reforms, there is now a realistic possibility that Ireland will be one of the first countries to witness industrial action over the use of artificial intelligence in education. 

Even if this impasse is eventually resolved, this foregrounds the genuine need for a discussion of what role we want AI to play in our education system. At the moment, this debate in education is polarised and it’s the worst type of polarised debate – one where both sides are correct.

On one side, many teachers, rightly, are concerned about the potential for students to cheat with AI and the completely impossible responsibility placed on them to detect this. On the other, supporters of the reforms, also quite rightly, can point to the benefits of project-based assessment and how experience of using AI will be an increasingly valuable life-skill.

Between the hype and the handwringing, there's a very basic truth about AI in education: it’s not the Messiah, but it's also not simply a very naughty bot.

What's your view on this issue?

You can tell us here

For a real discussion of AI in education, we need three ingredients: the first is an honest examination of the problems we have in our education system; the second is an informed discussion about what AI can and can’t do; and the third is the real involvement of teachers and students.

This should all start from a principle of 'Responsible AI'. In academia, it’s difficult to move without some conference, grant proposal or job-ad mentioning 'Responsible AI'. 

Though this has diluted the meaning of quite a complex idea, there’s a straightforward way to think about it: AI should solve problems we actually have and not problems that Big Tech, or AI itself, thinks we might have.

Grind schools

So, let’s start with a glaring problem in our education system in Ireland: grind schools. These schools dominate our education system, both financially and mentally. Financially, this "shadow education" was estimated by the ESRI at €60m. 

Mentally, our discourse is dominated by their notes, study tips and predictions (though I’ve yet to see an audit on these predictions). This is all the more flagrant during exam season. 

There are several problems with these schools and there’s a need for increased regulation, especially in the online space: one online school was recently caught peddling outdated material as a predicted paper and editing their own reviews.

The education gap

Beyond this, however, we have to ask if we really want an education system where a school, The Institute of Education, can sell for close to €130m? This is in a context where other schools are faced with a need to employ unqualified teachers (13,540 at post-primary and primary level). 

Indeed, the shortage of teaching staff between advantaged and disadvantaged schools in Ireland is one of the biggest in OECD countries. This is a social and a political problem, which needs to be addressed at a structural level. 

However, AI can make a contribution here: by making some tasks more efficient for teachers and by spreading educational resources to all students.

Teachers and AI

Three months ago, I launched Pulc.AI: a free essay corrector for the Leaving Certificate. We now have 1500 teachers and 3000 Leaving Cert students on the platform. 

As an academic, I’ve seen how disastrous AI can be in the classroom when let loose. Yet that doesn’t mean we can’t find a place for AI; when built with teachers, it can help improve education, especially in under-resourced schools.

In the last three months, I’ve been in contact with well-over 100 teachers. These conversations, almost universally, featured two emotions: confusion and enthusiasm.

Confusion at the current status of AI in schools; uncertainty as to how they were expected to deal with AI; and stress at the fact that this seemed to be all down to the individual teacher to determine when, how and why AI should be used. 

Quite rightly, these concerns are part of the basis for ASTI’s rejection of the Leaving Certificate reforms.

But what has stood out more than the confusion is the genuine enthusiasm from teachers to build responsible AI for their classrooms. What the debate around the Leaving Cert reforms on AI might mask is the real interest from teachers in these tools. 

From the newly qualified to those with 30+ years under the belt, I now get emails every day which show a real curiosity to learn more about this technology and an investment in making it better.

Dr Cillian O Fathaigh: 'AI should solve problems we actually have and not problems that Big Tech, or AI itself, thinks we might have.'

Importantly, this makes the real role of AI in education clear: it is a tool. It’s a very powerful tool, but it is not a replacement for the inspiration and personal connection of teachers. It won’t solve all the structural problems of education inequality, but it can help make teachers’ and students’ lives better.

This enthusiasm points to a need for our discussion around AI and the Leaving Cert reforms to go beyond settings limits and rules. We also need to find ways for teachers and students to positively engage with AI and to be involved in actively shaping these tools. Responsible AI, in the end, won’t build itself.

• Dr Cillian Ó Fathaigh is is an Assistant Professor in Philosophy at the Jagiellonian University Kraków and a Visiting Research Fellow at King’s College London. He currently leads a survey of AI teaching across 7 European universities. In March, he co-created www.pulc.ai: a Leaving Cert essay corrector and AI exam coach.