Understanding Young Adults’ Purchase Intentions for AI-driven Advertisements Using Explainable Machine Learning Models

This study investigates the determinants of young adults’ purchase intentions for artificial intelligence (AI)-driven ads and tests the predictive capabilities of machine learning models using XAI methods. A total of 324 participants aged 18–26 completed the survey from September 2024 to February 2025. The study utilised established scales for purchase intention, cultural influence, AI perception, technology access and cognitive responses. A random forest classifier was used to make predictions, and its performance was tested on more than 20 behavioural, cultural and technology-based indicators, using 300 estimators and a maximum depth of 10. Stratified 5-fold cross-validation was used to assess the model’s performance, with an average accuracy of 66.05%, indicating moderate predictive power. However, the model’s performance varied across intention-level classes, possibly due to class imbalance. The model correctly predicted the class ‘Neutral to Positive Intention’. According to SHapley Additive exPlanations-Local Interpretable Model-agnostic Explanations results, prior purchase intention was the primary predictor and technology access, cultural attitudes, societal perceptions, confusion and collectivism had minimal impact on the prediction.