DEVELOPING HUMANISTIC DESIGN METRICS FOR AI-MEDIATED MENTORSHIP IN COMPULSORY AND CONTINUING EDUCATION FRAMEWORKS

The addition of AI-mediated mentorship into mandatory (K-12) and lifelong learning has the potential of being scalable individualized, but existing systems are based on algorithmic performance rather than humanistic attributes like empathy, cultural sensitivity, and equity to learners. Thisdisconnects results in diminished motivation, higher dropout rates (up to 35% in underserved contexts), and inequitable outcomes, particularly in diverse regions like India, where digital divides persist. In this regard, suggest a detailed research questionnaires framework to conduct and testhumanistic design measurements to fit AI mentorshipframeworks. The model conceptualizes six main measures, which are empathy responsiveness, cultural adaptability, equity access, motivational alignment, ethical transparency and learner agency in a 45-item Likert-scale tool (5-8 itemsin each measure). Instrument development followed a systematic process: theoretical grounding in humanistic pedagogy, expert Delphi validation (content validity ratio >0.70), and pilot testing (n=50). 250 participants (n=130 in compulsory education in schools in Tamil Nadu; n=100 in continuingeducation programs in Coimbatore) in a quasi-experimental study, resulting in 230 analyzed cases. Pre/post tests were used to measure the humanistic AI prototypes and the conventional AI baselines. Exploratory factor analysis extracted six factors explaining 72.4% variance (KMO=0.89), withconfirmatory fit (CFI=0.96, RMSEA=0.04) and high reliability (Cronbach’s α=0.87-0.92). Key results include 28% gains in overall satisfaction (t=5.62, p<0.001), 22% in agency (t=4.89, p<0.001), and metric-specific uplifts (e.g., 32% equity in compulsory settings).

Hali tarjima qilinmagan