Prince George’s County Board of Education and Board of Community College Trustees for Prince George’s County – Vendor Registration PG 503–26
HB0589 would create new civil and criminal liability rules for artificial intelligence software that is capable of causing personal injury or death. Under the bill, a person who intentionally, knowingly, or negligently designs or creates such software would be strictly liable for damages if the software is used to cause injury or death, and could also face a civil penalty tied to Maryland’s noneconomic damages cap. The bill also adds a new criminal offense in the Criminal Law Article for designing or creating AI software with that capability.
The criminal penalties vary by mental state. Intentional conduct would be a felony punishable by up to 20 years in prison and/or a $100,000 fine; knowing conduct would be a misdemeanor punishable by up to 10 years and/or a $50,000 fine; and negligent conduct would be a misdemeanor punishable by up to 5 years and/or a $25,000 fine. The bill would take effect October 1, 2025, and would amend both the Courts and Judicial Proceedings Article and the Criminal Law Article of the Maryland Code.
If enacted, the bill would create a new subtitle in the Courts and Judicial Proceedings Article governing personal injury or death caused by implementation of artificial intelligence software, and a new criminal prohibition in the Criminal Law Article. It would expose AI developers and designers to strict civil liability and significant criminal penalties based on how they designed or created software capable of causing physical harm or death, potentially affecting software developers, AI companies, product designers, and others involved in AI system creation. The bill would also interact with Maryland’s existing noneconomic damages framework by tying the civil penalty to the applicable cap under § 3–2A–09.
There is little recorded committee or floor discussion available for this bill, and no votes are listed in the provided materials. The bill was ultimately withdrawn by the sponsor in the House, which suggests it did not advance through the legislative process. Based on the text alone, the measure appears to reflect a strong precautionary approach toward AI-related harm, but the absence of recorded debate makes it difficult to assess broader legislative sentiment beyond the sponsor’s initial proposal.
The main point of contention is the bill’s broad and severe treatment of AI design and creation, especially its inclusion of negligence as a basis for criminal liability and strict civil liability. Critics would likely focus on the difficulty of defining software “with the capability of causing personal injury or death,” the potential chilling effect on AI research and development, and whether criminal penalties are appropriate for design decisions that may be indirect or uncertain in their causal chain. Supporters would likely emphasize public safety, accountability, and the need to deter dangerous AI systems before they cause harm.