Biden's AI Executive Order Achieves First Major Milestones (AI EO January Update) — AI: The Washington Report – Mintz

On January 29, 2024, the White House published a fact sheet detailing the actions that have been taken pursuant to President Joe Biden’s October 30, 2023 executive order on “Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence” (“AI EO”). As covered in our timeline of the executive order, a number of provisions were scheduled to come into effect by the January 28, 2024 deadline. According to the White House’s announcement, more than two dozen actions pursuant to the executive order have been completed so far.
In this week’s newsletter, we will cover five major provisions enacted since October 2023. Our key takeaways are:
 
            Actions pursuant to the AI EO completed as of January 29, 2024. (Source)
 
Over the past few years, experts have warned that powerful AI tools could pose “potentially catastrophic effects on society” if not properly regulated. For instance, rogue autonomous systems “could facilitate the creation of novel bioweapons and lower barriers to obtaining such agents.” AI models that have the capacity to show “high levels of performance at tasks that pose a serious risk to security, national economic security, national public health or safety” are referred to in the AI EO as “dual-use foundation models.”[1]
Pursuant to the Defense Production Act, the AI EO directs the Secretary of Commerce to require companies “developing or demonstrating an intent to develop potential dual-use foundation models to provide the Federal Government, on an ongoing basis, with information, reports, or records” related to the following information:
The White House’s January 2024 fact sheet on the implementation of the AI EO announced that the Defense Production Act has been invoked to “compel developers of the most powerful AI systems to report vital information, especially AI safety test results, to the Department of Commerce.”[4] As Ben Buchanan, a White House special adviser on AI put it in a statement to the press, regulatory authorities want “to know AI systems are safe before they’re released to the public — the president has been very clear that companies need to meet that bar.”
On January 29, 2024, the Commerce Department published a proposed rule concerning US-based Infrastructure as a Service (IaaS) providers and their foreign resellers entitled, “Taking Additional Steps To Address the National Emergency With Respect to Significant Malicious Cyber-Enabled Activities.”
The Commerce Department promulgated the proposed rule pursuant to the AI EO’s requirement that the Commerce Secretary propose “regulations that require United States IaaS Providers to submit a report to the Secretary of Commerce when a foreign person transacts with that United States IaaS Provider to train a large AI model with potential capabilities that could be used in malicious cyber-enabled activity,” among other requirements.
IaaS providers allow customers to access computing resources, usually over the cloud. As discussed in our newsletter on AI and chips, cloud computing services allow firms that do not have access to sufficient computing resources to develop their own powerful AI models. The proposed Commerce Department rule is intended to ensure that malign entities cannot circumvent US trade restrictions, principally on powerful chips, to develop advanced AI models.
The proposed rule would require that US-based IaaS providers and their foreign resellers (“providers”) “verify the identity of foreign customers.” Providers would be required to “report to the Department [of Commerce] when they have knowledge they will engage or have engaged in a transaction with a foreign person that could allow that foreign person to train a large AI model with potential capabilities that could be used in malicious cyber-enabled activity.”
The Secretary of Commerce would be granted the power to “prohibit or impose conditions on the opening or maintaining with any U.S. IaaS provider of an Account, including a Reseller Account, by any foreign person located in a foreign jurisdiction found to have any significant number of foreign persons offering U.S. IaaS products used for malicious cyber-enabled activities, or by any U.S. IaaS provider of U.S. IaaS products for or on behalf of a foreign person.”
This rule can be understood as an escalation of the United States’ strategic competition with the People’s Republic of China (PRC) regarding the development of advanced technologies. If implemented as written, this rule would complement restrictions on the sale of advanced semiconductors to the PRC, limiting options for PRC firms to access the hardware requisite to develop leading-edge AI models.
Comments on this rule can be submitted through the Federal Register’s online portal. The comment period for this proposed rule closes on April 29, 2024.
A major accomplishment of the AI EO has been the National Science Foundation’s (NSF) launch of the National AI Research Resource (NAIRR) pilot.
In June 2020, Representative Anna Eshoo (D-CA-16) introduced the National AI Research Resource Task Force Act, a bill that would establish a task force to “investigate the feasibility and advisability of establishing a national artificial intelligence research resource” and “propose a roadmap detailing how such resource should be established and sustained.” Congress folded the bill into the National Defense Authorization Act of 2021, and by June 2021, the NAIRR Task Force had formally launched.
For over two years, the NAIRR Task Force constructed a proposal for the NAIRR. Their work culminated in their January 2023 final report. The report forcefully advocated for the NAIRR, stating that the resource “would transform the U.S. national AI research ecosystem and facilitate the ability to address societal-level problems by strengthening and democratizing participation in foundational, use-inspired, and translational AI R&D in the United States.” In the months following the issuance of the report, prominent AI firms and research institutions voiced their support for the NAIRR Task Force’s proposal.
In response to these developments, the AI EO orders the Director of the NSF to “launch a pilot program implementing the [NAIRR], consistent with past recommendations of the NAIRR Task Force.” On January 24, 2024, the NSF fulfilled this requirement by formally announcing the launch of the NAIRR pilot. At the time of writing, NAIRR has eleven government partners, including the Departments of Energy (DOE) and Defense (DoD), the National Institutes of Health (NIH), and the Defense Advanced Research Projects Agency (DARPA).
The pilot also has 25 private sector, nonprofit, and philanthropic partners, including the largest digital technology and AI firms in the United States. The NSF’s announcement states that the NAIRR pilot is seeking additional private sector and nonprofit partners and that interested parties should email NAIRR for further details.
The NAIRR pilot will have four primary initiatives:
As of January 2024, the NAIRR pilot has launched two programs.
On January 24, 2024, the NSF and DOE announced a research call to all “meritorious advanced computing proposals by US-based researchers and educators.” Selected projects will receive access to computational resources provided by NAIRR partners, such as the DOE’s Oak Ridge National Laboratory Summit supercomputing system.
Researchers must submit an application detailing a project “focused on advancing Safe, Secure and Trustworthy AI,” such as an intervention for “assuring that model functionality aligns with societal values and obeys safety guarantees.”
Applicants have until 8:00 pm EDT on March 1, 2024, to submit their applications. For more information, please visit the research call web portal.
The NAIRR pilot has compiled “government and non-government contributed resources aligned with the NAIRR Pilot goals, such as pre-trained models, AI-ready datasets, and relevant platforms.” At the time of writing, the pilot resources page lists over a dozen resources from NAIRR partners such as DoD, DOE, and NIH. More resources are expected to be added in the coming months.
As discussed in our newsletter on the AI EO, the order explicitly encourages the FTC to consider whether to exercise its rulemaking authority “to ensure fair competition in the AI marketplace and to ensure that consumers and workers are protected from harms that may be enabled by the use of AI.” Pursuant to this order, on December 20, 2023, the FTC announced a rulemaking concerning the Children’s Online Privacy Protection Act (COPPA). COPPA is a 1998 law empowering the FTC to regulate the collection of children’s data by online services.
The proposed December 2023 rule would expand the scope of practices barred under COPPA, limit exceptions, and subject entities participating in Safe Harbor programs to greater scrutiny. “By requiring firms to better safeguard kids’ data, our proposal places affirmative obligations on service providers and prohibits them from outsourcing their responsibilities to parents,” commented FTC Chair Lina Khan in reference to the proposed rule. If instituted, this proposed rule would constitute the first change to the COPPA Rule since 2013.
Provisions of the proposed rule include:
The comment period for the proposed rule closes on March 11, 2024. Comments can be submitted through the Federal Register’s online portal.
To build administrative capacity to address the challenges posed by AI, the AI EO calls for a cross-agency hiring surge to increase “AI talent in the Federal Government.” According to the White House’s January 2024 fact sheet on the implementation of the AI EO, the “Office of Personnel Management has granted flexible hiring authorities for federal agencies to hire AI talent, including direct hire authorities and excepted service authorities.” Additionally, existing government-wide tech talent programs such as Presidential Innovations Fellows and the US Digital Corps have scaled up hiring for AI talent in 2024 across “high-priority AI projects.”
Interested applicants should visit the federal government’s AI talent hub to explore job openings and other relevant opportunities.
While January 2024 saw the completion of many actions ordered by the AI EO, the most consequential deadline, April 27, 2024, is still on the horizon. There are 30 tasks due on the April deadline alone, not to mention the almost a dozen actions due in late February and March. These actions implicate agencies across the federal bureaucracy and touch on a wide swath of topics, including copyright, civil rights, national defense, and immigration. While January has been a busy month in Washington, as far as federal actions on AI go, the year is just getting started.
We will continue to monitor, analyze, and issue reports on these developments. Please feel free to contact us if you have questions as to current practices or how to proceed.
 
 
Subscribe To Viewpoints
Get our latest thought leadership
Review Cookie Management Center
© 2024 Mintz, Levin, Cohn, Ferris, Glovsky and Popeo, P.C. All Rights Reserved.

source

Leave a Reply

The Future Is A.I. !
To top
en_USEnglish