New Delhi: Excelsoft Technologies and UK-based awarding body AQA have agreed to form a joint AI taskforce to research how artificial intelligence can be used responsibly in marking handwritten student responses. The move comes as both organisations acknowledge that human examiners will continue to lead high-stakes marking, even as AI tools evolve.
The taskforce will combine AQA’s experience in large-scale assessments with Excelsoft’s capabilities in AI, data and e-assessment systems. According to both organisations, the collaboration aims to examine how AI can assist examiners in improving consistency and efficiency without replacing human judgement. The initiative focuses on developing models that support, rather than automate, decision-making in high-stakes examinations.
Key areas of work include testing human-in-the-loop workflows, evaluating bias-aware approaches for diverse student groups, and designing secure data pipelines for handwritten scripts. The group also plans to develop explainability frameworks so that AI-assisted marking processes can be clearly documented and governed. Ethical guidelines specific to exam-marking contexts will also be drafted as part of the research.
Excelsoft Chairman and Managing Director Sudhanva Dhananjaya said the work seeks to ensure that any AI used in marking maintains fairness, transparency and data protection. AQA Chief Assessment Technology Officer Justin Coombs noted that AI may become a useful support tool, provided its deployment remains accountable and subject to rigorous testing. Both organisations emphasised that the objective is to build confidence among students, schools and regulators as AI tools become more common in assessment systems.
The taskforce will include data scientists, psychometricians, AI engineers, examiners, assessment designers and ethicists from both sides. Its findings may inform future product design, assessment policies and industry discussions, with outputs to be shared through white papers or pilots where appropriate. The initiative reflects the sector’s growing interest in examining the practical and ethical boundaries of AI in high-stakes education settings.
