A complex landscape of state laws surrounding artificial intelligence (AI) is posing challenges for businesses across the United States, with each state taking its own approach to regulating this rapidly evolving technology.
In Utah, lawmakers are deliberating over legislation that would mandate certain businesses to disclose whether their products interact with consumers autonomously, without human intervention.
Meanwhile, in Connecticut, proposed legislation aims to impose stricter regulations on transparency regarding the inner workings of AI systems deemed “high risk.”
These states are part of a group of 30, along with the District of Columbia, that have either introduced or enacted laws imposing various constraints on the design and use of AI systems.
These laws address a wide range of issues, including child protection, data transparency, bias reduction, and consumer safeguarding in sectors such as healthcare, housing, and employment.
Goli Mahdavi, a lawyer with Bryan Cave Leighton Paisner, describes the current regulatory landscape as “a mess for business,” citing the uncertainty caused by the evolving bills and statutes.
The absence of direct federal regulation stems from a lack of consensus among US lawmakers on the necessity of such laws to govern AI.
In contrast, other parts of the world, such as the European Union and China, have implemented comprehensive AI regulations.
However, despite the lack of federal action, state laws in the US do reflect priorities set by the federal government. President Biden’s executive order from last October focused on the responsible application of AI, while a subsequent directive requires developers to disclose safety test results.
While state laws share common themes, subtle differences can complicate compliance for businesses.
For instance, while several states have adopted consumer protection laws requiring notification and opt-out provisions for automated decision-making, the specifics vary.
Some states also require risk assessments for AI use, while others mandate bias audits or restrict AI applications in specific sectors.
The passage of these laws has been facilitated by the prevalence of single-party control in many state legislatures, with roughly 40 states dominated by a single party last year.
In this evolving regulatory landscape, businesses operating in multiple states must navigate a patchwork of laws to ensure compliance with varying requirements regarding AI design and usage.
