The Biden administration released an executive order on artificial intelligence on October 30th, setting new privacy, anti-discrimination, and safety standards for AI throughout the executive branch, while also seeking to accelerate innovation and bolster the AI workforce. Officially titled the “Executive Order on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence,” it’s informally being called the AI EO. The White House also released a Fact Sheet, providing a high level summary of the different actions within the EO.
The order is organized around a number of themes, with specific directions to federal agencies in each theme:
- New Standards for AI Safety and Security
- Protecting Americans’ Privacy
- Advancing Equity and Civil Rights
- Standing Up for Consumers, Patients and Students
- Supporting Workers
- Promoting Innovation and Competition
- Advancing American Leadership Abroad and,
- Ensuring Responsible and Effective Government Use of AI
Much of the EO is centered around setting up regulations for tech companies that want to sell programs and software to federal agencies. By leveraging the federal government as a customer for these products, using the Defense Production Act, the Biden Administration is attempting to force these companies to adopted good best practices and, by extension, influence what is offered in the wider marketplace, both within the US and globally.
However, this order goes far beyond federal bureaucratic procurement requirements, and has several items that directly and indirectly impact the nation’s research community and enterprise. Section 5, “Promoting Innovation and Competition” is where the majority of the interest for the computing community is. One of the biggest items is proposing that NSF pilot the National AI Research Resource (NAIRR), “a tool that will provide AI researchers and students access to key AI resources and data,” according to the order’s fact sheet. According to the wording in the order itself, NAIRR will, “pursue the infrastructure, governance mechanisms, and user interfaces to pilot an initial integration of distributed computational, data, model, and training resources to be made available to the research community in support of AI-related research and development.” NAIRR has a lot of bipartisan support in Congress (see more details here), and is likely to be included in any near-term legislation covering AI; but getting the program off the ground is likely to run headlong into the difficult funding situation in Congress.
There’s a lot of other good news for the computing community in this section of the EO: “fund and launch at least one NSF Regional Innovation Engine that prioritizes AI-related work;” “establish at least four new National AI Research Institutes, in addition to the 25 currently funded as of the date of this order;” and “support activities involving high-performance and data-intensive computing,” at DOE and in coordination with NSF to, “establish a pilot program to enhance existing successful training programs for scientists, with the goal of training 500 new researchers by 2025 capable of meeting the rising demand for AI talent.” There is also direction to the federal agencies to expand AI research grants in areas like healthcare and climate change. All of that is good news for the research community.
Included in Section 5 are provisions covering rules for high-skilled immigration. Unfortunately, much of it is streamlining bureaucratic immigration and visa practices, and promoting what the Biden Administration has already done in this area. The fact is the Biden Administration did about as much as they could do on high-skilled immigration in January of 2022. To do more will require Congress to take up the issue and pass legislation on general immigration policy, something that has not happened in a generation.
In Section 9, “Protecting Privacy,” NSF, in partnership with DOE, is directed to establish a new Research Coordination Network (RCN) to, “promote the adoption of leading-edge privacy-preserving technologies by federal agencies.” The RCN is meant to, “enable privacy researchers to share information, coordinate and collaborate in research, and develop standards for the privacy-research community.” NSF is further directed to, “where feasible and appropriate, prioritize research — including efforts to translate research discoveries into practical applications — that encourage the adoption of leading-edge PETs solutions for agencies’ use.” Privacy-protecting technologies are mentioned prominently throughout the EO.
This executive order is quite expansive and covers a lot of areas. And it’s just the first one in a likely set of actions from the Administration covering AI. There is reference to a forthcoming national security memorandum on, “AI and security.” And the order’s Fact Sheet concludes with, “(m)ore action will be required, and the Administration will continue to work with Congress to pursue bipartisan legislation to help America lead the way in responsible innovation.” This will not be the last word from the Biden Administration on the topic of AI. CRA will continue to tack these actions and will report out to the community; please be sure to check back for future updates.