Colman Noctor: Is AI in the classroom a welcome revolution or a shortcut to nowhere?

Irish Examiner

a day ago

As artificial intelligence threatens to reshape how children learn, we must balance its promise of personalised teaching with the risk of losing what lies at the heart of education — learning through human connection. With ChatGPT-5 and other AI programmes freely available, our understanding of teaching and learning could be transformed.
Many parents are fascinated by, and uneasy about, AI's impact on their children's education. While AI offers efficiency and personalisation, it also raises concerns about fairness and human interaction.
Educators and parents need to understand this change because our goal must be to ensure that children learn effectively, not just quickly.
AI's greatest strength is its ability to adapt to each user. Adaptive learning systems can recognise a learner's struggles, motivations, and pace. A 2020 review of 37 studies by researcher Shuai Wang, at the University of Illinois, found that 86% were positive about learning from AI technologies. Many learning platforms are adopting personalised AI approaches, and tools like language education app Duolingo can tailor lessons to individual needs.
Teachers also benefit from the rapid technological changes. AI promises to handle repetitive tasks, such as grading quizzes, creating assignments, and monitoring attendance, freeing them to provide one-to-one support and relational teaching.
The potential of AI to support children with learning needs is of particular interest to me as my teenage daughter is dyslexic. I am interested in exploring how word-prediction software and voice-enabled interfaces can help students with dyslexia or speech difficulties keep pace with their peers.
Risks of AI in education
Every benefit has a corresponding risk, and when it comes to AI in education, the main one is over-reliance and the erosion of real-life connections. Screens and algorithms cannot replace the warmth, empathy, and spontaneity of teachers, who are crucial for social-emotional development.
Bias is another concern. AI often inherits or amplifies biases from its training data, which could potentially disadvantage minorities and reinforce stereotypes. We have already seen this happen with social media algorithms, which have contributed to the fragmentation and polarisation of societal views.
Academic integrity is also under threat. Universities report widespread use of generative AI, raising fears about the loss of independent thinking. AI detection tools were once seen as a safeguard, but newer systems can now 'humanise' text to evade detection.
The capacity for AI to write essays and complete assignments undetected is particularly troubling in professional training programmes, such as medicine or nursing. To be safe practitioners, these students must fully understand the course content. If AI undermines academic standards and evades detection, the consequences could be significant. Suppose a student submits an AI-generated essay on which medications should not be taken together, but does not know what it contains; this lack of understanding poses risks for competency in the clinical field.
While there are established safeguards to ensure clinical competency through practice placements, concerns persist regarding students' comprehensive knowledge of the course content.
AI could also deepen digital inequality. Without access to high-speed internet or the latest tools, some students risk falling even further behind.
Growing use among students
Many third-level students are aware of AI's limitations. A 2025 University of Florida survey found that students value AI's instant feedback and study support, but worry about accuracy, the loss of critical thinking, data privacy, and bias.
Regardless, the use of ChatGPT among students is on the rise, particularly in secondary schools. In the US, 26% of teens (13–17 years old) reported using ChatGPT for schoolwork in 2024, which is double the 2023 figure. They found that older secondary students were most likely to use it, and when asked what they deemed 'acceptable use', 54% said for research purposes, dropping to 29% for solving maths problems.
So how can parents help children learn in a world increasingly being driven by AI?
1. Encourage discernment: Let children use AI tools like ChatGPT or personalised tutoring, but teach them to question: Where did this answer come from? Could it be wrong? How would you explain it yourself? These questions are essential to developing critical-thinking skills.
2. Champion human connection: Young people need to see teachers, mentors, and parents who model empathy, resilience, and humour. These are essential life qualities that AI cannot replicate.
3. Prioritise AI literacy: Teach children that AI can be wrong, biased, or unclear. Discussions about privacy, data, and ethics have never been more critical, and they need to take place at home, not just be left to teachers in school. To raise the issue of ethical and responsible use, you could say: 'AI can give quick answers, but it doesn't always get things right. How do you check whether the information is true before using it?' Or, 'It's fine to use AI to brainstorm ideas, but your teachers will want to see your thinking. How do you decide when to stop relying on the tool and write in your own words?'
4. Value effort alongside efficiency: In a world where AI can write essays or instantly solve difficult maths questions, we should at least require that students produce their own initial draft. If AI provides a solution to a problem, we could ask the student to explain the process to confirm they genuinely grasp the concept.
5. Keep learning yourself: The more you understand AI's strengths and weaknesses, the better you can guide children through this fast-changing world.
Keeping humans at the centre
The Higher Education Policy Institute recently concluded that while AI can support teachers, it cannot, and must not, replace them. Critical thinking, creativity, and emotional connection are essential human qualities.
UNESCO, the UN agency that promotes education, science, culture, and communication, also acknowledges AI's potential to make education more inclusive and equitable, but only if we remain vigilant about its ethical implications, privacy, and fairness. The agency argues that AI should not be used as an answer generator, but as a tutor. Rather than prompting it to 'give me the answer to this question', attempt the problem and tell ChatGPT to 'show me where I went wrong'.
The next two years will be transformative for the education sector. While a change to the current system is needed, we must be careful not to throw the baby out with the bathwater when it comes to educational progress and hoped-for efficiency.
We could learn from Estonia, which has historically invested heavily in digital infrastructure and is the highest PISA-ranked country for global recognition of technological learning.
The Estonian Department of Education has partnered with OpenAI and other platforms to provide training for teachers and establish guardrails for data protection, as well as strategies to encourage critical thinking.
The Estonian government seems to see the value of AI in education, but understands the necessity to be prepared, so that it can be integrated in a manner that optimises its benefits and minimises the risks.
From an Irish perspective, the Department of Education committed to developing guidelines for schools regarding the integration of AI in April 2024, but this is still in development.
AI promises significant benefits, and can positively transform our future. However, we need to engage with it in a way that considers the benefits and the risks. If we allow AI to dominate our educational interactions, we risk creating a superficial education system where efficiency takes precedence over understanding.
As we approach the return to school, let's consider how we leverage AI's benefits without losing sight of the human heart of learning.
Dr Colman Noctor is a child psychotherapist