Re: Notice of Public Comment Period, Office of Management and Budget; Advancing Governance, Innovation, and Risk Management for Agency Use of Artificial Intelligence (88 Fed. Reg. 75,625-75,626, November 3, 2023)
Dear Director Young:
The U.S. Chamber appreciates the opportunity to provide comment to the Office of Management and Budget (OMB) on its “Advancing Governance, Innovation, and Risk Management for Agency Use of Artificial Intelligence” (“Draft Guidance”).1 Furthermore, we appreciate OMB’s acknowledgment that “AI is improving operations and efficiency across the federal government2.” Public sector utilization of artificial intelligence can improve the efficiency of the federal government and facilitate easier public interaction. Therefore, it is imperative that it should be a goal of OMB to facilitate federal government utilization of AI technology, to promote IT modernization efforts throughout the Federal Government.
At the outset, we wish to reiterate our continued concerns with the short comment period provided. Accordingly, the U.S. Chamber, with other associations,3 sent a letter on November 14 emphasizing that the “complexity of the issues involved and other concurrent and overlapping federal regulatory requests require extending the comment period to allow for more thoughtful and substantive feedback.” Because this request for additional time was no granted, the Chamber, and other stakeholders, can only provide limited feedback and comment.
The Chamber strongly supports the Draft Guidance’s emphasis on cross-agency and public- private collaboration. This is vital and can help agencies leverage multiple stakeholders' collective resources, expertise, and innovation to address complex and interrelated policy challenges. Furthermore, consistency and harmonization between agencies and between existing regulatory regimes, where appropriate, can speed up the adoption of critical AI tools.
Furthermore, these efforts must be based upon and aligned with existing laws, legal standards, and consumer protections. Therefore, the Draft Guidance should consider if existing controls on the federal government’s use of information technology already govern AI-specific risks.
We recommend the following issues should be addressed within the Draft Guidance:
- Strengthening AI Governance: Annual postings of all AI use cases may not be sustainable or practicable. As a result, the draft guidance would fail to achieve the intended transparency goals. Furthermore, this requirement may unintentionally create an undue burden for agencies. We feel further discussions, studies, and standards development are necessary regarding model notices to determine what should be reported in the public inventory. We would also like to stress that such a scope of inventory should only be for those uses considered high-risk.
- Managing risks from the use of AI: The Chamber believes it is essential to align the Draft Guidance with existing Federal Acquisition Regulations (FAR), and OMB should solicit public comments to achieve that goal.
- Definition of “Rights-impacting:" The definition of “rights-impacting” could capture non-high risk AI utilization. A risk-based approach is vital to prioritize the allocation of resources for systems with heightened risk. Without this approach, the federal government could inadvertently suppress IT modernization adoption within the federal government. We would further emphasize the need to align the Draft Guidance with the National Institute of Standards and Technology risk-management framework. Furthermore, it is important to note that there are neither technical nor professional standards for algorithmic audits, and inadequate standards may skew risk requirements that fail to identify and address intended issues. Finally, we would encourage the federal government to engage with all stakeholders regarding “negative feedback” regarding specific AI systems to ensure that such concerns have merit.
- Waivers for minimum practices: Regarding waivers that can be granted, we strongly encourage OMB to add language so such waivers can be issued to protect organizations' exposure to sensitive or other protected information.
- Definitions: The Chamber strongly supports the alignment of regulatory terms to industry-specific AI definitions and terminology to ensure non-AI technologies are appropriately carved out of AI definitions and oversight. Where possible, the guidance should reflect internationally accepted definitions endorsed by the U.S., such as those promulgated by the Organization for Economic Cooperation and Development. The following terms should receive ample feedback from stakeholders before being finalized, “rights-impacting” and “potential harms to works wellbeing.” Clarity of terms and definitions is essential, so we welcome further opportunities to discuss industry- supported definitions, including terms like “third-party vendor” that will be crucial in implementing any policies resulting from the guidance.
- Timelines: Currently, no timelines are tied to federal agencies’ impact assessment, testing, or independent evaluations. A lack of clarity around specific timelines could lead to potential bottlenecks in AI utilization within the federal government, wherein agencies will be waiting on one antecedent step prior to making procurement decisions.
In conclusion, we appreciate the opportunity to provide high-level feedback on the Draft Guidance. The Chamber believes that more time is necessary for stakeholders to review and provide more substantive feedback, as the Draft Guidance will substantially impact the government’s ability to take advantage of AI tools. We encourage OMB to provide other opportunities to receive input from stakeholders regarding these issues. The Chamber stands willing and ready to work with OMB to ensure that all AI the government utilizes is done safely and securely.
Sincerely,
Michael Richards
Director, Policy
Chamber Technology Engagement Center
3. https://americaninnovators.com/advocacy/letter-to-omb-on-ai-draft-guidance-extension/
