Bill > HB4098

This bill requires artificial intelligence (AI) businesses that use or sell "high risk AI systems" to conduct thorough risk assessments before deployment, evaluating potential discrimination, bias, safety risks to children and vulnerable populations, privacy concerns, and catastrophic risks. These businesses must then submit a transparency report to the Arizona Attorney General detailing the AI system's purpose, operation, risk assessment results, and bias mitigation measures, and this report must also be publicly posted on the company's website. The bill mandates that AI businesses implement bias mitigation strategies and conduct quarterly audits to identify and correct any discriminatory or unsafe AI behavior. If a high-risk AI system might be accessed by children, specific parental monitoring and content reporting tools are required. The Attorney General is empowered to enforce these provisions, review transparency reports, and impose civil penalties of up to $50,000 per violation, with reduced penalties possible for inadvertent violations if the business complies with recognized standards and the Children's Online Privacy Protection Act. The bill defines "artificial intelligence business," "frontier model" (a complex AI model with high training costs or derived from another frontier model), "high risk AI system" (one impacting children, healthcare, employment, public safety, or education, or likely to cause discrimination or privacy violations), and "large developer" (an entity that has trained frontier models with significant financial investment, excluding academic research).

Business and Industry

https://apps.azleg.gov/BillStatus/BillOverview/85620