In this article, we delve into the transformative potential of AI-power in education, the impact on exams being one aspect, and explore the implications for the future of education.Only 95% Accurate!During a recent gathering at the FE News Conference in Birmingham, a representative from NCFE unveiled a groundbreaking development in education assessment. We were told that with the advent of Artificial Intelligence (AI), according to their research, exams are now being marked with an astounding accuracy of 95 percent. The exam-marking algorithm has been trained on a huge dataset of student responses to past papers and has access to all the course study materials and recommended reading. It’s a top student!This revelation signifies a huge leap forward in the integration of AI within education, promising to revolutionise the way we evaluate student performance. The potential implications of this advancement are far-reaching, as it offers the promise of not only streamlining the assessment process but also new opportunities for personalised learning experiences.In this article, we delve into the transformative potential of AI-power in education, the impact on exams being one aspect, and explore the implications for the future of education.The Accuracy Will Only ImproveNCFE has decided not to use AI for marking their exams just yet; the reason given is trust, or rather, lack of it. People won’t like to have their exam marked by a computer and potentially, to rub salt into the wounds, fail them. Research confirms that 52% of people are more ‘concerned’ than ‘excited’ by AI. Why’s that?The current lack of trust is understandable as this shift challenges long-standing assumptions about human involvement in these processes. The idea that a machine can generate and evaluate questions with such high accuracy is a testament to the rapid advancement of technology, and it’s natural to feel surprised or even apprehensive about such a change. Don’t know about you, but I am surprised!Considering the revolution in written assessment, there are also some other significant challenges to overcome to build accountability and trust. AI systems struggle with ‘explainability‘, that is, to articulate how they arrive at specific marks or decisions, unlike human markers. Then, there is unreliability, where slight input variations can result in significant output differences, and bias, arising from the data AI systems are trained on, reflecting wider societal attitudes.Even if we accept the 5% inaccuracy, over time that inaccuracy, along with the other issues, will be eroded to zero as the programming is generative, meaning that the performance improves in time. From a human marker perspective, 5% inaccuracy isn’t too bad and who’s to say the one ‘inaccurately marked’ paper in twenty is by the AI? Having been a marker of exams myself, I’m in favour of using the algorithm; it means there is no reliance on the potential incosistency between markers, no human errors (hey, enough about me!) and we’ll be happy with the myriad benefits of consistency, accuracy, fairness in marking, no unpredictable biases, little cost and instant feedback and results. Bingo!As the improvements accumulate, it’s likely that AI will become more integrated into the education system, and examining boards will adapt their communication strategies. Transparency about the use of AI in marking is important for maintaining trust and ensuring fairness, as will be the inevitable ‘regulatory’ processes of quality-assurance and expert interventions.Not everyone agrees with that perspective. David Lindsay, a history teacher at St. Ambrose School in Manchester, believes:“Humanities and social science subjects will present many problems because they require subjective judgements that only human examiners can supply. AI may be better suited to sciences , languages and maths. Teacher marking and assessment is an important part of the teacher-pupil relationship as is knowledge and understanding of...

Ver todos los episodios