Civil Actions - Product Liability - Artificial Intelligence Systems
HB 712 creates a new Maryland product-liability framework specifically for artificial intelligence systems. It establishes a new subtitle in the Courts and Judicial Proceedings Article allowing lawsuits against AI developers for harm caused by a dangerous or defective product, including claims based on defective design, failure to provide adequate instructions or warnings, and breach of express warranty. The bill also allows a deployer to be sued in place of a developer in certain circumstances, such as when the deployer materially and substantially changes the system or intentionally misuses it contrary to an express warranty.
The bill defines key terms such as artificial intelligence system, high-impact AI system, generative AI system, consequential decision, deployer, developer, harm, and synthetic relationship. It excludes AI used only for peer-reviewed scientific research and exempts certain small deployers and open-source users from the developer definition. The bill also creates rebuttable presumptions that an AI system is not dangerous or defective if the developer followed documented testing, auditing, mitigation, and disclosure practices, or if the deployer implemented and followed a reasonable risk-management policy. It further authorizes the Attorney General to bring enforcement actions seeking injunctions, civil penalties, restitution, damages, and other relief.
HB 712 would affect Maryland tort law by adding a new statutory cause of action and clarifying how liability is allocated in AI-related injury cases. It states that contributory negligence is not a complete defense, though a plaintiff’s recovery is reduced by the plaintiff’s percentage of fault, and it permits joint and several liability between developers and deployers. The bill also provides that autonomous action by the AI system is not a defense, and it expressly supplements existing common-law and statutory product-liability remedies rather than replacing them.
The overall sentiment reflected in the bill text is strongly supportive of consumer protection and risk management, with the preamble emphasizing the need to protect Marylanders from harms caused by AI systems. Because there are no committee transcripts or recorded votes provided, there is no direct evidence of legislative debate or opposition in the available materials. Based on the structure of the bill, the measure appears designed to balance innovation with accountability by giving developers and deployers compliance-based defenses while expanding remedies for harmed individuals and state enforcement.
Notable points of potential contention include the breadth of the definitions, especially what qualifies as a high-impact AI system and a consequential decision, and the extent of liability imposed on developers and deployers. The bill’s treatment of open-source AI, small businesses, minors, and synthetic relationships may also draw scrutiny, as may the rebuttable presumptions tied to industry best practices and risk-management policies. Another likely issue is the bill’s decision to limit contributory negligence as a defense while still allowing fault-based reduction of damages, which could be debated by both consumer advocates and technology industry stakeholders.
HB 712 would add a new subtitle to the Maryland Courts and Judicial Proceedings Article creating a statutory AI product-liability regime. It would expand state law by authorizing private civil actions and Attorney General enforcement against AI developers and deployers, establishing standards for defect, warning, warranty, and liability allocation, and supplementing existing common-law product-liability claims rather than displacing them. The bill would also affect developers, deployers, insurers, and users of high-impact and generative AI systems, particularly where the systems are used in consequential decisions, critical infrastructure, vehicles, medical devices, or other high-risk settings.
The bill’s tone and findings are strongly pro-regulation and pro-consumer, reflecting concern that AI systems can cause measurable harm and should be subject to clearer legal accountability. No committee testimony or vote history is provided, so there is no recorded evidence of support or opposition from legislators or stakeholders in the available materials. On its face, the bill seeks to create a balanced framework that protects the public while offering compliance-based protections for developers and deployers that follow testing, disclosure, and risk-management practices.
Likely points of contention include whether the bill’s definitions are too broad, especially the scope of “high-impact artificial intelligence system,” “consequential decision,” and “synthetic relationship.” Technology companies and deployers may object to expanded liability, joint and several liability, and the Attorney General’s enforcement authority, while consumer advocates may argue the bill does not go far enough if the rebuttable presumptions are too easy to satisfy. The carveouts for small businesses, open-source users, and peer-reviewed research, along with the special treatment of minors and the limits on contributory negligence, are also likely to be debated.