HB1331 creates a new Maryland consumer-protection framework for artificial intelligence systems that are used in high-stakes decisions. It defines “high-risk artificial intelligence systems” as systems that make, or substantially factor into, decisions producing legal or similarly significant effects for consumers, and it defines “algorithmic discrimination” as differential treatment that negatively affects a person based on protected characteristics such as race, sex, age, disability, religion, national origin, veteran status, and others. The bill requires developers and deployers of such systems to take reasonable precautions against foreseeable discriminatory harms, with separate duties for developers who sell these systems and deployers who use them.
For developers, the bill requires standardized disclosures to purchasers, public-facing summaries of the systems offered for sale, and notice to purchasers and the Attorney General if the developer learns a system is causing or likely to cause algorithmic discrimination. For deployers, the bill requires a risk management policy, annual impact assessments, consumer-facing disclosures, and a process allowing consumers to correct data and appeal adverse decisions, including human review in many cases. The bill also creates a limited small-deployer exception and allows appeals to be declined when an appeal would not be in the consumer’s best interest, such as where delay could create a safety risk.
The bill also adds a separate subtitle governing agreements involving voice and likeness clones. It makes certain agreements to perform personal or professional services unenforceable if they allow creation or use of a digital copy of a person’s voice or image without clear, understandable terms about intended uses and licensing, unless the person was represented by counsel or by a labor union under a collective bargaining agreement covering digital copies. The bill further amends Maryland’s unfair trade practices law so that violations of the new AI subtitle are treated as unfair, abusive, or deceptive trade practices enforceable under the Maryland Consumer Protection Act.
Its impact on state law is significant because it would impose new compliance, disclosure, documentation, and recordkeeping obligations on AI developers and deployers operating in Maryland, while also giving the Attorney General oversight authority and enforcement leverage. It would also create new consumer rights around AI-assisted decisions and new contract rules for AI-generated voice and likeness replicas. The bill amends the Commercial Law Article and relies on the State Finance and Procurement Article’s definition of artificial intelligence.
The available context shows no recorded votes or committee testimony, so there is no documented public debate in the materials provided. Based on the bill text, the overall tone is precautionary and consumer-protective, with the legislation aiming to reduce discrimination and increase transparency rather than restrict AI use outright. Likely points of contention include the compliance burden on businesses, the breadth of disclosure and assessment requirements, the scope of the small-deployer exemption, and the carve-outs for trade secrets and security-sensitive information.
HB1331 would add a new subtitle to the Commercial Law Article regulating developers and deployers of high-risk AI systems, and it would make violations of those requirements an unfair, abusive, or deceptive trade practice under Maryland consumer-protection law. It would also create new contract enforceability rules for agreements involving digital copies of a person’s voice or likeness, and it would rely on the State Finance and Procurement Article’s definition of artificial intelligence. The bill would therefore expand state oversight of AI-related consumer decisions, increase disclosure and documentation obligations, and give the Attorney General enforcement authority.
The bill’s apparent sentiment is broadly supportive of consumer protection, transparency, and anti-discrimination safeguards in AI use. The text is structured to require risk management, notice, and human review rather than banning AI systems. Because no committee transcript or vote record is provided, there is no documented opposition or support in the supplied materials, but the design of the bill suggests a policy preference for cautious deployment of AI in consequential consumer settings.
Likely areas of contention include whether the bill’s reporting, disclosure, and annual assessment requirements are too burdensome for developers and deployers, especially smaller businesses; whether the definition of high-risk AI is broad enough to capture too many systems; and whether the consumer appeal and human-review requirements are operationally feasible in time-sensitive contexts. Another likely issue is the balance between transparency and protection of trade secrets, security-sensitive information, and information already protected by law. The voice-and-likeness clone provisions may also draw concern from entertainment, media, and labor stakeholders over contract enforceability and licensing practices.