Fifth Flash Report on AI Provides New Guardrails for the House

On April 17, 2024, the Committee on House Administration (CHA) released its latest AI Flash Report. The ten-page report announces the establishment of AI-guardrails for the House, provides updates on the use of generative artificial intelligence (GenAI) technology across the Legislative branch, and announces the creation of a new Chief Administrative Office (CAO) AI Coordinator and AI Center of Excellence.

This fifth AI flash report continues the subcommittee’s dedication of providing transparent updates about how the House of Representatives is adapting to the emergence of GenAI.

The report explores CHA’s actions of hosting Member-led discussions about AI’s influence on the internal operations of the Legislative branch, including a full committee hearing on AI innovation in the Legislative branch that took place in January, as well as a March roundtable on ensuring safe and appropriate use of AI within the institution. From these informative discussions, the Committee has moved forward to issue AI guardrails for the House (available in detail on page 3 of the report), with the intent for them to be applied to any AI tool or technology used within the House.

CHA’s newly established AI guardrails for the House are: 

  • human oversight and decision-making

  • clear and comprehensive policies

  • robust testing and evaluation

  • transparency and disclosure

  • education and upskilling

The committee believes that the roundtable it hosted is “the first-known instance of elected officials directly discussing AI’s use in parliamentary operations” and that the AI guardrails are the first of their kind to be issued by a legislative body in the world.

In addition to the updates on actions undertaken by CHA, Wednesday’s Flash Report also provides updates on the investment in improved governance structures, use cases, and staff upskilling pursued by the Government Publishing Office (GPO), Library of Congress (LOC), Smithsonian Institution, Architect of the Capitol (AOC), US Capitol Police (USCP), and House Chief Administrative Office (CAO).

Since the first flash report released in September of 2023, the committee has requested the Legislative branch agencies under its oversight jurisdiction to submit quarterly reports detailing internal management efforts to responsibly update protocols. A few highlights from last quarter’s reports (summarized in the Flash Report) include: 

  • The formation and kick-off of GPO’s AI Governance Committee as well as GPO’s announcement of three pilot programs. One pilot involves the creation of a GPO intranet chatbot to assist employees with more quickly accessing information to aid with customer inquiries.

  • The LOC’s and Smithsonian’s engagement in co-chairing an international working group to focus on how AI will impact libraries, archives, and museums. 

  • The continued establishment and updating of Legislative branch AI use-case inventories (many of which are publicly accessible and linked within the report). 

  • The AOC IT Council’s focus on crafting and successful adoption of AI governance and policy recommendations for all AOC employees to follow.

  • The USCP’s initiative to develop an AI charter to inform the agency’s future approach to integration of AI technologies. 

  • The CAO’s continued investment in the creation of a comprehensive AI policy, as well as future plans to establish the position of AI Coordinator and an AI Center of Excellence. 

Through the updates provided in the quarterly AI Flash Reports, CHA continues to ensure that the House’s internal response to AI is both proactive and transparent, placing the institution at the forefront of legislatures around the world in adapting to this emerging technology. This recent report demonstrates how it takes a village to get a legislative institution to adapt to a new paradigm, but that CHA is doing vital work in coordinating all the power players to make sure the institution’s response to this emerging technology is anything but passive.

Editor’s Note: The AI guardrails roundtable hosted by the Committee on House Administration is believed to be the first of its kind to bring together elected officials to candidly discuss a legislative institution’s internal guidance to this emerging technology. As such, and due to the success of the roundtable, committee staff have contributed a resource guide to the Bússola Tech LegisTech Library allowing other parliaments the ability to explore echoing the model. This contribution to the international network is notable and is a practice that the House and Senate should explore more often.

Previous
Previous

AI Tools for Casework: A Guide for Congressional Staff

Next
Next

Public Witness Testimony on Strengthening Congressional Casework to House Appropriations Legislative Branch Subcommittee for FY25