Ohio labor groups back bill to limit AI in state benefit decisions

Labor pushes for human oversight

Organized labor in Ohio is advocating for House Bill 1032, a piece of legislation introduced by Representatives Tristan Rader and Christine Cockley. The bill aims to establish guardrails for how state agencies utilize artificial intelligence, particularly regarding employment and taxpayer-funded benefits. As reported by Hoodline, the measure seeks to ensure that humans remain at the forefront of decisions that impact people's livelihoods.

The proposed law would require state employers to be transparent about their AI policies and disclose how they are currently using the technology. It specifically targets the use of automated systems in high-stakes scenarios where a person's economic stability is at risk. By mandating disclosure, the bill intends to pull back the curtain on how algorithms might be influencing hiring, promotions, or disciplinary actions.

Labor representatives express concern that workers are often unaware of how these tools are being deployed behind the scenes. Brian Pearson, a representative for the North Shore AFL-CIO, noted that workers worry about AI replacing them or making life-altering decisions without human context. He argued that automated models should not replace the compassion and understanding that human workers provide when evaluating unique personal circumstances.

Restricting automated benefits

A central component of House Bill 1032 is the restriction on how state agencies distribute public resources. The bill would specifically prohibit the use of AI to determine whether an individual qualifies for benefits through Job and Family Services. This means that while AI might assist in some administrative tasks, it cannot be the final arbiter of eligibility for essential state support.

This push for human intervention comes as the Ohio Department of Job and Family Services has already begun integrating AI into its operations. The agency recently rolled out four different AI tools and received an award for incorporating the technology into its unemployment services. However, the department has already built a requirement for human verification of any AI-generated output into its current strategy.

The debate reflects a broader national conversation about the ethics of automated governance in the public sector. While some see AI as a way to streamline massive amounts of data, others fear it removes the nuance required for social services. By banning AI from making final benefit decisions, the bill seeks to preserve a human connection in the distribution of taxpayer funds.

Ohio's existing AI framework

Ohio is not starting from a blank slate regarding the governance of artificial intelligence within its government. The state currently operates a multi-agency AI Council designed to provide statewide oversight for the use of generative AI solutions. This council works alongside existing administrative policies to ensure that new technologies are monitored as they are adopted by various departments.

Under the state's current Administrative Policy IT-17, any new use case for AI must follow a specific approval path. This process requires that the technology be defined and routed through agency development teams and executive leadership before reaching the AI Council. These rules are intended to provide a structured way for the state to manage the risks and benefits of emerging software.

Despite these existing structures, lawmakers are still grappling with how to effectively regulate the rapidly evolving field. Representative Thaddeus Claggett, who chairs the House Technology and Innovation Committee, acknowledged that predicting the effects of AI legislation is difficult. He noted that regulating the technology will require continued work rather than a single, simple solution to cover every possible scenario.

The legislative landscape

The timing of House Bill 1032 may pose a significant hurdle for its passage during the current legislative session. Representative Claggett suggested that it is unlikely for a bill introduced this late in the year to pass before the General Assembly concludes. This sentiment reflects a common trend where complex technology bills face a difficult path to becoming law in a short timeframe.

On a national scale, the push for AI regulation in the workplace has been widespread this year. Analysis shows that states considered 155 different bills addressing AI in the workplace, with themes like transparency, accountability, and worker privacy leading the way. However, many of these proposals died in committee without ever reaching a final vote in either chamber.

Other states have seen varying levels of success in passing similar measures. Colorado and Connecticut enacted laws requiring employers to notify applicants about automated hiring tools and explain any adverse decisions. In contrast, other efforts, such as a New York bill requiring large businesses to report on AI's effect on hiring, failed to be signed into law, illustrating the complexity of the current regulatory environment.