According to the Legislative Budget Board (LBB), HB 5118 would have an estimated negative impact of $1,215,000 on General Revenue Related Funds for the biennium ending August 31, 2027. The bill would not itself make an appropriation, but the LBB states that it could provide the legal basis for an appropriation to implement the bill.
The projected cost is limited to fiscal year 2026. The LBB estimates a $1,215,000 one-time cost from the General Revenue Fund in fiscal year 2026, with no additional state fiscal impact shown for fiscal years 2027 through 2030. The cost is tied to the Texas Workforce Commission’s need to hire a contractor to conduct the required study on artificial intelligence systems and automated employment decision tools in employer and state agency hiring practices by the bill’s deadline.
The Department of Information Resources is not expected to have a significant fiscal impact from assisting with the study. The LBB also reports that the Texas Workforce Commission does not anticipate information technology expenditures, and no significant fiscal implication to units of local government is anticipated.
Texas Policy Research recommends that lawmakers vote NO on HB 5118. Although the Committee Substitute is narrower than the introduced version, it still expands the role of state government by directing the Texas Workforce Commission, in consultation with the Department of Information Resources, to study the use of artificial intelligence systems and automated employment decision tools in hiring by both state agencies and private employers. The bill is framed as a study, but its required scope goes beyond neutral information gathering and directs the agency to examine oversight deficiencies, auditability, regulatory protections, vendor disclosure requirements, accountability requirements, external independent review, and future legislative or other action.
The bill grows the size and scope of government by moving the Texas Workforce Commission into a new policy area: state evaluation of artificial intelligence tools used in employment decisions. The bill also grants the commission rulemaking authority necessary to administer the Act and implement related programs authorized under it. Even though the Act expires September 1, 2027, the required report is designed to produce recommendations for future legislative or administrative action, including a proposal to establish a method of funding research on artificial intelligence bias. That structure creates a practical pathway for continued state involvement after the study ends.
The bill increases the burden on taxpayers. The LBB estimates a negative impact of $1,215,000 to General Revenue Related Funds through the 2026–27 biennium, with the cost occurring in fiscal year 2026. The stated cost driver is the Texas Workforce Commission’s anticipated need to hire a contractor to conduct the study by the bill’s deadline. The bill does not make an appropriation, but it could provide the legal basis for one. For a bill that does not directly remedy a defined statutory failure, this represents a meaningful taxpayer cost to fund a study that may later be used to justify additional spending or regulation.
The Committee Substitute does not immediately impose new compliance duties on private employers in the way the introduced version did. The introduced bill would have required employers using automated employment decision tools to obtain independent bias audits, post audit results publicly, notify applicants, and disclose how the tool would evaluate them. The substitute removes those direct mandates and replaces them with a study. However, the substitute still reaches private employers by requiring the state to study employer use of these tools and to evaluate transparency, disclosure, independent review, accountability, and regulatory protections. That creates a foreseeable risk of future regulatory burdens on businesses, especially employers and technology vendors.
From a free-enterprise and limited-government perspective, the bill is concerning because it appears to be a precursor to regulation rather than a narrowly tailored review of state agency practices. Existing state and federal antidiscrimination laws already govern unlawful employment discrimination. Before creating a taxpayer-funded study aimed at identifying regulatory gaps and future oversight mechanisms, the legislature should require clear evidence that existing law is inadequate. The bill instead directs a state agency to search for deficiencies and propose future action, which risks turning a speculative concern into a justification for expanding government authority.
The bill expands government scope, imposes a taxpayer cost, creates rulemaking and programmatic authority, and lays the groundwork for future regulation of private-sector hiring technology. A narrower proposal limited strictly to internal state agency use of artificial intelligence in hiring, with no authority to study or recommend regulation of private employers and no new taxpayer-funded research framework, would present a materially different policy question.