AI for research assessment: opportunities and challenges from India

Introduction. Artificial intelligence (AI) is beginning to reshape the infrastructures of scholarly communication, raising important questions about trust and transparency. This study examines researchers’ perspectives on AI-driven scientific replication prediction tools as emerging components of research assessment. Method. Qualitative, in-depth, semi-structured interviews were conducted with 19 faculty and doctoral scholars in India to explore how researchers perceive and engage with the scientific replication AI tool. Analysis. Our analysis focused on how information practices, institutional incentives, and sociocultural contexts shape the adoption of AI technologies in scholarly work. Interview transcripts were studied using thematic analysis using a collaborative and iterative coding process. Results. Participants mention that limited funding, infrastructure, and access to advanced tools or high-quality datasets make replicating studies difficult in India. They recognised the value of AI tools for surfacing reproducibility assessments during literature reviews and study design. They advocated for hybrid human–AI systems that balance the efficiency of automation with the nuanced judgment of experts. Conclusion. This study situates AI replication tools within broader scholarly infrastructures, highlighting the need for design features that enhance explainability, fairness, and cultural sensitivity. It also represents the challenges faced by Indian researchers when it comes to replication and open science.