The Mexican Senate has initiated an Artificial Intelligence Law that classifies violations into three tiers — minor, serious, and "very serious" — with the last category carrying potential prison sentences, according to the proposal. "Very serious" offenses include mass surveillance without legal basis, deployment of lethal autonomous systems without human oversight, and deliberate cognitive or political manipulation via AI. The bill would establish a National AI Authority and a National Fund for AI Development and employs a risk-based framework with heightened obligations for high-impact systems.