Resolution to explore artificial intelligence in staff work

Summary#

SJR-17 was a Utah Senate joint resolution, not a bill amending an existing Act. It failed, so it did not change the law or direct legislative staff to use artificial intelligence (AI). Its stated goal was to encourage staff to explore AI and related technologies to improve legislative information and analysis.

  • The resolution encouraged legislative staff to explore, evaluate and responsibly use AI to improve the clarity and accessibility of legislative information and analyses.
  • It said AI use should preserve human oversight, protect confidential information, maintain analytical integrity, and comply with applicable legal, ethical and security standards.
  • It encouraged staff training, best practices and governance measures, but did not set specific requirements, timelines or enforcement.

What it means for you#

  • Legislative staff and lawmakers: Because the resolution failed, it creates no new duties or authority. If it had passed, it would have encouraged staff to explore AI use; it did not require staff to use AI or create a process for reviewing their work.
  • Public: The resolution did not change public access to legislative information or guarantee changes to its accuracy, clarity or availability.

Money#

The resolution appropriated no money; no cost estimate is provided.

  • It encouraged training and governance measures, but the materials do not say whether these would cost money or how they would be funded.

What is unclear#

  • The resolution did not define AI or related technologies.
  • It did not specify which tasks or information could be handled with AI, or how staff would protect confidential information in practice.
  • It did not set rules for checking AI-generated analysis, assigning responsibility for errors, or measuring whether the tools improve accuracy or accessibility.
  • It did not specify costs, a start date or oversight. Since the resolution failed, none of its proposed steps took effect.

Case for#

  • The resolution identifies growing demands on staff from the volume and complexity of legislation. AI tools could help with research, drafting, summarization, analysis and data processing.
  • A possible argument for the approach is that exploring tools while emphasizing human oversight and confidentiality could help staff improve legislative materials without treating AI output as a substitute for staff judgment.
  • More accessible legislative information could help the public understand the Legislature’s work; the resolution states this as a goal, not a guaranteed result.

Case against#

  • One concern is that the resolution encourages AI use but does not set concrete rules for checking accuracy, protecting confidential information or deciding who is responsible for mistakes.
  • It encourages training and governance measures without identifying who would provide them or what they might cost.
  • It is unclear how the Legislature would judge whether AI use actually improves the timeliness, consistency, clarity or accessibility of its information.

Votes

The recorded vote on this bill, member by member.

  1. Division No. 1652799

    Passed4 members voted

    Senate Comm - Favorable Recommendation

    FOR · 3AGAINST · 1

    Each dot is one member. Hover or tap one to see who.

    3For

    • Republican2
    • Democratic1

    1Against

    • Republican1

    5 members were paired and sat out the vote.