AI-Driven Counselling Interventions for Predicting and Preventing Xenophobic Attacks Among South African Youths

Abstract

This study examined AI-driven counselling interventions for predicting and preventing xenophobic attacks among South African youths. Specifically, it investigated how artificial intelligence–supported counselling strategies identify early warning signs of xenophobic tendencies, reduce hostile attitudes, and promote social cohesion. A descriptive survey design with a correlational approach was adopted. The population comprised youths aged 18–35 years from selected provinces in South Africa with histories of xenophobic incidents. A sample of 500 respondents was selected using multistage sampling techniques. Data were collected using three instruments: the AI-Driven Counselling and Xenophobia Prevention Questionnaire (AICXPQ), AI Predictive Risk Assessment Checklist (AIPRAC), and Focus Group Discussion Guide (FGDG). Validity was established through expert review, while reliability coefficients of 0.87 and 0.81 were obtained. Data were analyzed using descriptive statistics, Pearson correlation, multiple regression, and ANOVA at 0.05 level of significance. Findings revealed a significant relationship between AI-driven counselling and prediction of xenophobic attacks (r = .62), and between AI-assisted counselling and reduction of xenophobic attitudes (r = .68). AI-based early detection systems significantly predicted prevention of xenophobic violence (R² = .50), while social cohesion differed significantly based on exposure levels (F = 14.36, p < .05). The study concludes that AI-driven counselling enhances prevention and social outcomes and recommends its ethical integration into counselling practice and youth development policies.