OMB’s newly released guidance makes strides towards more efficient and effective government use of AI, as the focus now shifts to strengthening transparency and procurement guidance
(WASHINGTON)—Today, the Center for Democracy & Technology (CDT) welcomed the Office of Management and Budget’s (OMB) release of its final guidance on governance and risk management for federal agencies’ use of artificial intelligence (AI). The guidance’s stated purpose is to direct agencies in innovating and improving their AI governance, while managing risks stemming from AI uses that impact the safety and rights of individuals.
“This guidance is an important step for the federal government to lead by example in the responsible use of AI,” says CDT President and CEO Alexandra Reeve Givens. “Rather than have individual agencies grapple with governing and mitigating risks of AI on their own – when many agencies are trying to solve similar problems in these areas – we are glad to see detailed guidance to support agencies’ responsible and transparent use of this technology.”
This final guidance follows OMB’s solicitation of comments on their draft memorandum for agency use of AI, to which CDT and a number of other civil society organizations contributed in early December 2023. The final OMB AI memo responds to that comment process in several key ways, in that it:
- Improves transparency by committing to providing consistent guidance for how agencies should report their AI inventories and requires agencies to release additional details, such as whether AI uses are rights- or safety-impacting; additional information on how related risk management practices are implemented; which AI use cases are granted waivers from required minimum risk practices; and why an AI tool is the best solution for a particular use case, rather than a non-AI alternative;
- Ensures that agencies’ design, development, and use of AI is informed by consultation with affected communities, including underserved communities, the public, civil society, and the federal workforce on the risks of issues like individual use cases; how those risks are managed; whether AI is meeting the agency’s equity goals; applicable opt-out procedures; and sharing best practices across agencies; and
- Guides agencies’ evaluation and use of AI-driven tools that are provided by third-party companies, including detailed guidance on transparency and how to hold vendors accountable for their performance as well as stronger safeguards if a federal agency seeks to purchase technology services that utilize biometrics and facial recognition technology.
“We are heartened to see that OMB took feedback from the comment process seriously, and made a number of improvements to increase transparency and support consistent processes across agencies. Importantly, OMB also elaborated on how federal agencies can responsibly procure AI. This stronger guidance, along with OMB’s new request for information on further procurement guidelines, plays a vital role in informing whether and how public dollars are spent on AI. But this guidance is only a first step towards the most important aspect of the process – implementation,” Givens concluded.
Despite the improvements OMB made in the final guidance released today, CDT identified missed opportunities where OMB could have gone further to:
- Identify additional minimum risk practices, such as data minimization, to ensure AI systems are safely and effectively serving individuals; and
- Establish a redress process if a chief AI officer inappropriately grants a waiver from the minimum risk practices, as there is no recourse to challenge the validity of the decision to exempt AI uses.
Given previous expectations set through federal actions, and that deployed government AI tools are currently making consequential decisions for individuals, CDT encourages federal agencies to act now to rigorously implement this guidance, ahead of the December 1, 2024, deadline.
***
The Center for Democracy & Technology (CDT) is the leading nonpartisan, nonprofit organization fighting to advance civil rights and civil liberties in the digital age. We shape technology policy, governance, and design with a focus on equity and democratic values. Established in 1994, CDT has been a trusted advocate for digital rights since the earliest days of the internet. The organization is headquartered in Washington, D.C., and has a Europe Office in Brussels, Belgium.