NJ bill proposes $5 million penalty for AI firm violations

New Jersey Assembly Democrats have proposed legislation that would impose strict liability on owners, operators, and developers of generative AI platforms. If enacted, the measure would hold these entities financially responsible for crimes committed by their AI products or agents, marking a shift in how state law treats emerging technology.

The bill, referred to by sponsors as an accountability act, includes a provision for civil penalties of up to $5 million per violation. This amount is in addition to compensatory and punitive damages. Assemblyman Andrew Macurdy, a Democrat from Union, serves as the primary sponsor. He stated that current product liability laws are insufficient because cases involving AI companies often take a long time to resolve in court. Macurdy described the proposed legislation as the next level beyond existing statutes.

The Assembly’s Science, Innovation and Technology Committee held a hearing on the issue at the State House in Trenton earlier this week. No votes were cast during the session, but testimony was provided by national and state experts. Beverly Brown Ruggia, financial justice program director at New Jersey Citizen Action, argued that state policymakers must act in the absence of meaningful federal regulation. Ruggia noted that society already understands the definitions of fraud, theft, and murder, and that someone must be held accountable when AI commits these acts.

Rob Eleveld, cofounder of the Transparency Coalition, testified that approaching AI regulation through product liability law is the most effective way to change corporate behavior. He predicted that as many as 10 states could adopt similar legislation within the next year. Eleveld said New Jersey could be the first state in the nation to enact such laws if lawmakers move quickly.

Opposition has emerged from the New Jersey Business & Industry Association. Jack Ramirez, a representative for the group, raised concerns about how the proposed penalties might impact smaller and emerging AI companies operating in the state. While Ramirez acknowledged the inclusion of an affirmative defense clause for companies making reasonable efforts to prevent violations, he urged lawmakers to define those efforts more clearly. He also raised concerns about other bills in the package, including measures requiring labels on AI-generated content and annual risk disclosures to the state Attorney General.

Chris Tully, a Democrat from Bergen and chair of the committee, stated that the committee will act on the measures. Tully noted that the public is looking for regulation and that the topic is of consequence. A poll conducted by Stockton University earlier this year found that 41% of registered voters believe increased AI use will make their lives worse, with a margin of error of plus or minus 3.9 percent.

The legislation follows a 2024 report by a special task force convened by former Gov. Phil Murphy, which highlighted the societal benefits of AI. Recent high-profile incidents, including a hack of the Hugging Face platform by autonomous agents from OpenAI, have elevated the urgency of the debate. Macurdy cited these events during the hearing, arguing that companies need financial incentives to proceed more carefully.