AI-Supported lecturer decision-making in higher education: a socio-technical perspective

Abstract Artificial intelligence (AI) has become an integral part of teaching and learning in higher education. Through its capacity to generate actionable insights from student data, AI systems provide new opportunities to support lecturers’ decision-making. However, it remains unclear how AI systems support and mediate lecturer decision-making across different decision types. Existing research has primarily focused on technical capabilities and learning outcomes, while paying limited attention to the socio-technical dynamics of AI-supported lecturer decision-making. To address this gap, this study develops an empirically grounded taxonomy of AI-supported lecturer decision-making and maps the relationships among decision types, AI systems, student data, and learning outcomes. Following PRISMA guidelines, a systematic literature review of 27 empirical studies published between 2016 and 2025 was conducted using qualitative content analysis. The findings identify eight lecturer decision types and show that AI support is concentrated in instructional, feedback, and assessment decisions, while emotional, ethical, administrative, curriculum, and learning-environment decisions remain underrepresented. Learning Analytics Dashboards were the most frequently reported system type, primarily supporting monitoring, feedback, and assessment functions based on textual and log data. From a socio-technical perspective, the AI systems described in this sample predominantly made behavioral student data visible and actionable for lecturers, and the decisions they supported were correspondingly concentrated in instructional, assessment, and feedback domains, while processes related to motivation, metacognition, emotions, and learning environments received far less support. These findings highlight an important gap: How can alternative forms of student data visibility, attentional guidance, and interface design support lecturer decisions?