Requires political advertisements generated substantially by AI to include disclaimers indicating AI involvement. Mandates specific formatting for AI disclaimers in graphic, audio, and video ads. Establishes fines for non-compliance and outlines exemptions for certain media and platforms.
Analysis summaries, actor details, and coverage mappings were LLM-classified and may contain errors.
This is a binding legislative act with mandatory disclosure requirements, specific enforcement mechanisms including civil fines, and criminal penalties for violations.
The document primarily addresses risks related to malicious actors using AI for disinformation and manipulation (4.1, 4.3), with minimal coverage of misinformation risks (3.1). It focuses specifically on political advertising contexts and does not substantially address other risk domains in the MIT taxonomy.
This document primarily governs the Information sector (broadcasting, telecommunications, digital platforms) and Professional and Technical Services (political consulting, advertising agencies). It also has limited application to Public Administration in the context of election-related communications.
The document focuses primarily on the Deploy and Operate and Monitor stages of AI-generated political advertisements. It does not address earlier lifecycle stages such as planning, data collection, or model development.
The document explicitly mentions artificial intelligence in the context of content generation for political advertisements. It does not define AI models, AI systems, or reference specific AI categories such as frontier AI, general purpose AI, or foundation models. No compute thresholds are mentioned.
Michigan State Legislature (The People of the State of Michigan)
The document is a legislative act enacted by the Michigan State Legislature, as indicated by the enacting clause and the formal legislative structure.
Michigan Secretary of State (rule-making authority), Michigan state courts (criminal and civil enforcement)
The Secretary of State is granted rule-making authority for implementation, while state courts enforce through criminal misdemeanor prosecution and civil infraction proceedings.
The document does not explicitly designate a monitoring body or establish monitoring mechanisms. Enforcement appears to be complaint-driven through the court system rather than proactive monitoring.
Persons, committees, or other entities that create, originally publish, or originally distribute qualified political advertisements; individuals paying for political advertisements
The law targets any person, committee, or entity that creates or distributes AI-generated political advertisements, as well as those who pay for such advertisements.
4 subdomains (3 Good, 1 Minimal)