This study investigates users' artificial intelligence (AI)-related competencies, including AI knowledge, skills, and attitudes, and identifies vulnerable user groups in the AI-shaped online news and entertainment environment. A survey of 1088 Dutch citizens identified five user groups through latent class analysis: average users, expert advocates, expert skeptics, unskilled skeptics, and neutral unskilled. The most vulnerable groups, characterized by low AI knowledge and skills, were older, less educated, and had lower privacy protection skills. The findings provide evidence for an emerging AI divide among users, similar to the digital divide, and highlight the need for education programs and explainable AI to address these disparities. The study also discusses the societal implications, emphasizing the importance of addressing the AI divide to ensure equitable access and outcomes for all users.This study investigates users' artificial intelligence (AI)-related competencies, including AI knowledge, skills, and attitudes, and identifies vulnerable user groups in the AI-shaped online news and entertainment environment. A survey of 1088 Dutch citizens identified five user groups through latent class analysis: average users, expert advocates, expert skeptics, unskilled skeptics, and neutral unskilled. The most vulnerable groups, characterized by low AI knowledge and skills, were older, less educated, and had lower privacy protection skills. The findings provide evidence for an emerging AI divide among users, similar to the digital divide, and highlight the need for education programs and explainable AI to address these disparities. The study also discusses the societal implications, emphasizing the importance of addressing the AI divide to ensure equitable access and outcomes for all users.