Reimagining Educational Assessment through Human–AI Collaborative Intelligence: A Conceptual Framework for Competency-Based Assessment in the AI Era

The rapid advancement of Artificial Intelligence (AI) has initiated a paradigm shift in educational assessment, offering unprecedented opportunities for personalization, continuous feedback, and evidence-driven instructional decision-making. However, contemporary discourse predominantly positions AI as a technological substitute for traditional assessment functions such as automated grading, adaptive testing, and predictive analytics. Such a technology-centric perspective inadequately captures the inherently human nature of educational assessment, which extends beyond measurement to encompass interpretation, contextual understanding, reasoning, and professional judgment. ethical generative AI, and educational measurement with the policy directions articulated in the National Education Policy (2020), the National Curriculum Framework for School Education (2023), and PARAKH, the article develops a comprehensive framework for responsible AI integration in Indian school education. Rather than advocating technological determinism, the paper proposes a balanced vision in which AI enhances teachers' capacity to understand learning more holistically while preserving equity, inclusion, transparency, and learner agency. The article concludes that the most transformative assessment systems of the future will not be defined by artificial intelligence alone but by the quality of collaboration between human intelligence and intelligent technologies. This paper argues that the future of educational assessment lies not in replacing teachers with intelligent algorithms but in fostering meaningful collaboration between human expertise and artificial intelligence. It proposes the Human–AI Collaborative Assessment (HACA) Framework, a conceptual model that reconceptualizes AI as an evidence amplifier rather than an autonomous evaluator. Within this framework, AI continuously gathers, synthesizes, and analyses multidimensional evidence of learning, while teachers retain the responsibility for interpreting evidence, contextualizing learner performance, and making pedagogically informed decisions. The paper further introduces the concept of an Assessment Evidence Ecosystem (AEE), which expands conventional assessment beyond test scores to include cognitive, behavioural, metacognitive, collaborative, reflective, and digital evidence generated across learning experiences. By integrating emerging scholarship on learning analytics, competency-based assessment,