The paper examines the normative and epistemic challenges of AI literacy with problems of applied ethics, open issues in the legal domain, and how they are related to digital divide concerns. The lack of skills, knowledge and understanding of how technology functions may depend on previous drivers of social inequality and digital division, and yet, today’s AI illiteracy can trigger a new digital divide much harsher than the first wave of the late 1990s and early 2000s. Threats of AI illiteracy are systemic, exponential, cumulative, and can be worsened by the lawmaker’s response, or lack thereof. The open issues of the digital divide are unpromising. Although LLMs hallucinations are only a part of a more intricated problem, it would be a first great step forward to increase the number of people capable of understanding they are just dealing with such hallucinations.