The White House Office of Science and Technology Policy (OSTP) has sought public input for the Biden administration’s National AI Strategy, acknowledging the potential benefits and risks of advanced AI. The Federation of American Scientists (FAS) was happy to recommend specific actions for federal agencies to safeguard Americans’ rights and safety. With U.S. companies creating powerful frontier AI models, the federal government must guide this technology’s growth toward public benefit and risk mitigation.

Recommendation 1: OSTP should work with a suitable agency to develop and implement a pre-deployment risk assessment protocol that applies to any frontier AI model.

Before launching a frontier AI system, developers must ensure safety, trustworthiness, and reliability through pre-deployment risk assessment. This protocol aims to thoroughly analyze potential risks and vulnerabilities in AI models before deployment. 

We advocate for increased funding towards the National Institute of Standards and Technology (NIST) to enhance its risk measurement capacity and develop robust benchmarks for AI model risk assessment. Building upon NIST’s AI Risk Management Framework (RMF) will standardize metrics for evaluation incorporating various cases such as open-source models, academic research, and fine-tuning of models which differ from larger labs like OpenAI’s GPT-4.

We propose the Federal Trade Commission (FTC), under Section 5 of the FTC Act, implement and enforce this pre-deployment risk assessment strategy. The FTC’s role to prevent unfair or deceptive practices in commerce is aligned with mitigating potential risks from AI systems.

Recommendation 2: Adherence to the appropriate risk management framework should be compulsory for any AI-related project that receives federal funding.

The U.S. government, as a significant funder of AI through contracts and grants, has both a responsibility and opportunity. Responsibility: to ensure that its AI applications meet a high bar for risk management.  Opportunity: to enhance a culture of safety in AI development more broadly. Adherence to a risk management framework should be a prerequisite for AI projects seeking federal funds.

Currently, voluntary guidelines such as NIST’s AI RMF exist, but we propose making these compulsory. Agencies should require contractors to document and verify the risk management practices in place for the contract. For agencies that do not have their own guidelines, the NIST AI RMF should be used. And the NSF should require documentation of the grantee’s compliance with the NIST AI RMF in grant applications for AI projects. This approach will ensure all federally funded AI initiatives maintain a high bar for risk management.

Recommendation 3: NSF should increase its funding for “trustworthy AI” R&D.

“Trustworthy AI” refers to AI systems that are reliable, safe, transparent, privacy-enhanced, and unbiased. While NSF is a key non-military funder of AI R&D in the U.S., our rough estimates indicate that its investment in fields promoting trustworthiness has remained relatively static, accounting for only 10-15% of all AI grants. Given its $800 million annual AI-related budget, we recommend that NSF direct a larger share of grants towards research in trustworthy AI.

To enable this shift, NSF could stimulate trustworthy AI research through specific solicitations; launch targeted programs in this area; and incorporate a “trustworthy AI” section in funding applications, prompting researchers to outline the trustworthiness of their projects. This would help evaluate AI project impacts and promote proposals with significant potential in trustworthy AI. Lastly, researchers could be requested or mandated to apply the NIST AI RMF during their studies.

Recommendation 4: FedRAMP should be broadened to cover AI applications contracted for by the federal government.

The Federal Risk and Authorization Management Program (FedRAMP) is a government-wide initiative that standardizes security protocols for cloud services. Given the rising utilization of AI services in federal operations, a similar system of security standards should apply to these services, since they are responsible for managing highly sensitive data related to national security and individual privacy.

Expanding FedRAMP’s mandate to include AI services is a logical next step in ensuring the secure integration of advanced technologies into federal operations. Applying a framework like FedRAMP to AI services would involve establishing robust security standards specific to AI, such as secure data handling, model transparency, and robustness against adversarial attacks. The expanded FedRAMP program would streamline AI integration into federal operations and avoid repetitive security assessments.

Recommendation 5: The Department of Homeland Security should establish an AI incidents database.

The Department of Homeland Security (DHS) needs to create a centralized AI Incidents Database, detailing AI-related breaches, failures and misuse across industries. Its existing authorization under the Homeland Security Act of 2002 makes DHS capable of this role. This database would increase understanding, mitigate risks, and build trust in AI systems’ safety and security.

Voluntary reporting from AI stakeholders should be encouraged while preserving data confidentiality. For effectiveness, anonymized or aggregated data should be shared with AI developers, researchers, and policymakers to better understand AI risks. DHS could use existing databases such as the one maintained by the Partnership on AI and Center for Security and Emerging Technologies, as well as adapt reporting methods from global initiatives like the Financial Services Information Sharing and Analysis Center.

Recommendation 6: OSTP should work with agencies to streamline the process of granting Interested Agency Waivers to AI researchers on J-1 visas.

The ongoing global competition in AI necessitates attracting and retaining a diverse, highly skilled talent pool. The US J-1 Exchange Visitor Program, often used by visiting researchers, requires some participants to return home for two years before applying for permanent residence.

Federal agencies can waive this requirement for certain individuals via an “Interested Government Agency” (IGA) request. Agencies should establish a transparent, predictable process for AI researchers to apply for such waivers. The OSTP should collaborate with agencies to streamline this process. Taking cues from the Department of Defense’s structured application process, including a dedicated webpage, application checklist, and sample sponsor letter, could prove highly beneficial for improving the transition of AI talent to permanent residency in the US.
Review the details of these proposals in our public comment.

The Federation of American Scientists (FAS) has identified several domains in the transportation and infrastructure space that retain a plethora of unsolved opportunities ripe for breakthrough innovation.

Transportation is not traditionally viewed as a research- and development-led field, with less than 0.7% of the U.S. Department of Transportation (DOT) annual budget dedicated to R&D activities. The majority of DOT’s R&D funds are disbursed by modal operating administrators mandated to execute on distinct funding priorities rather than a collective, integrated vision of transforming the nation’s infrastructure across 50 states and localities. 

Historically, a small percentage of these R&D funds have supported and developed promising, cross-cutting initiatives, such as the Federal Highway Administration’s Exploratory Advanced Research programs deploying artificial intelligence to better understand driver behavior and applying novel data integration techniques to enhance freight logistics. Yet, the scope of these programs has not been designed to scale discoveries into broad deployment, limiting the impact of innovation and technology in transforming transportation and infrastructure in the United States. 

As a result, transportation and infrastructure retain a plethora of unaddressed opportunities – from reducing the 40,000 annual vehicle-related fatalities, to improving freight logistics through ports, highways, and rail, to achieving a net zero carbon transportation system, to building infrastructure resilient to the impacts of climate change and severe weather. The reasons for these persistent challenges are numerous: low levels of federal R&D spending, fragmentation across state and local government, risk-averse procurement practices, sluggish commercial markets, and more. When innovations do emerge in this field, they suffer from two valleys of death: one to bring new ideas out of the lab into commercialization, and the second to bring successful deployments of those technologies to scale.

The United States needs a concerted national innovation pipeline designed to fill this gap, exploring early-stage, moonshot research while nurturing  breakthroughs from concept to deployment. An Advanced Research Projects Agency-Infrastructure would deliver on this mission. Modeled after the Defense Advanced Research Projects Agency (DARPA) and the Advanced Research Projects Agency-Energy (ARPA-E), the Advanced Research Projects Agency-Infrastructure (ARPA-I) will operate nimbly and with rigorous program management and deep technical expertise to tackle the biggest infrastructure  challenges and overcome entrenched market failures. Solutions would cut across traditional transportation modes (e.g. highways, rail, aviation, maritime, pipelines etc) and would include innovative new infrastructure technologies, materials, systems, capabilities, or processes. 

The list of domain areas below reflects priorities for DOT as well as areas where there is significant opportunity for breakthrough innovation:

Key Domain Areas

Metropolitan Safety

Despite progress made since 1975, dramatic reductions in roadway fatalities remain a core, persistent challenge. In 2021, an estimated 42,915 people were killed in motor vehicle crashes, with an estimated 31,785 people killed in the first nine months of 2022. The magnitude of this challenge is articulated in DOT’s most recent National Roadway Safety Strategy, a document that begins with a statement from Secretary Buttigieg: “The status quo is unacceptable, and it is preventable… Zero is the only acceptable number of deaths and serious injuries on our roadways.” 

Example topical areas include but are not limited to: urban roadway safety; advanced vehicle driver assistance systems; driver alcohol detection systems; vehicle design; street design; speeding and speed limits; and V2X (vehicle-to-everything) communications and networking technology.

Key Questions for Consideration:

• What steps can be taken to create safer urban mobility spaces for everyone, and what role can technology play in helping create the future we envision?
• What capabilities, systems, and datasets are we missing right now that would unlock more targeted safety interventions? 

Rural Safety

Rural communities possess their own unique safety challenges stemming from road design and signage, speed limits, and other factors; and data from the Federal Highway Administration shows that “while only 19% of the U.S. population lives in rural areas, 43% of all roadway fatalities occur on rural roads, and the fatality rate on rural roads is almost 2 times higher than on urban roads.”

Example topical areas include but are not limited to: improved information collection and management systems; design and evaluation tools for two-lane highways and other geometric design decisions; augmented visibility; mitigating or anti-rollover crash solutions; and enhanced emergency response. 

Key Questions for Consideration:

• How can rural-based safety solutions address the resource and implementation issues that are faced by local transportation agencies?
• How can existing innovations be leveraged to support the advancement of road safety in rural settings?

Resilient & Climate Prepared Infrastructure

Modern roads, bridges, and transportation are designed to withstand storms that, at the time of their construction, had a probability of occurring once in 100 years; today, climate change has made extreme weather events commonplace. In 2020 alone, the U.S. suffered 22 high-impact weather disasters that each cost over $1 billion in damages. When Hurricane Sandy hit New York City and New Jersey subways with a 14-foot storm surge, millions were left without their primary mode of transportation for a week. Meanwhile, rising sea levels are likely to impact both marine and air transportation, as 13 of the 47 largest U.S. airports have at least one runway within 12 feet of the current sea level. Additionally, the persistent presence of wildfires–which are burning an average of 7 million acres annually across the United States, more than double the average in the 1990s–dramatically reshapes the transportation network in acute ways and causes downstream damage through landslides, flooding, and other natural events.

These trends are likely to continue as climate change exacerbates the intensity and scope of these events. The Department of Transportation is well-positioned to introduce systems-level improvements to the resilience of our nation’s infrastructure.

Example topical areas include but are not limited to: High-performance long-life, advanced materials that increase resiliency and reduce maintenance and reconstruction needs, especially materials for roads, rail, and ports; nature-based protective strategies such as constructed marshes; novel designs for multi-modal hubs or other logistics/supply chain redundancy; efficient and dynamic mechanisms to optimize the relocation of transportation assets; intensive maintenance, preservation, prediction, and degradation analysis methods; and intelligent disaster-resilient infrastructure countermeasures. 

Key Questions for Consideration:

• How can we ensure that innovations in this domain yield processes and technologies that are flexible and adaptive enough to ward against future uncertainties related to climate-related disasters?
• How can we factor in the different climate resilience needs of both urban and rural communities?

Digital Infrastructure

Advancing the systems, tools, and capabilities for digital infrastructure to reflect and manage the built environment has the power to enable improved asset maintenance and operations across all levels of government, at scale. Advancements in this field would make using our infrastructure more seamless for transit, freight, pedestrians, and more. Increased data collection from or about vehicle movements, for example, enables user-friendly and demand-responsive traffic management, dynamic curb management for personal vehicles, transit and delivery transportation modes, congestion pricing, safety mapping and targeted interventions, and rail and port logistics. When data is accessible by local departments of transportation and municipalities, it can be harnessed to improve transportation operations and public safety through crash detection as well as to develop Smart Cities and Communities that utilize user-focused mobility services; connected and automated vehicles; electrification across transportation modes, and intelligent, sensor-based infrastructure to measure and manage age-old problems like potholes, air pollution, traffic, parking, and safety.

Example topical areas include but are not limited to: traffic management; curb management; congestion pricing; accessibility; mapping for safety; rail management; port logistics; and transportation system/electric grid coordination.

Key Questions for Consideration:

• How might we leverage data and data systems to radically improve mobility and our transportation system across all modes?

Expediting and Upgrading Construction Methods

Infrastructure projects are fraught with expensive delays and overrun budgets. In the United States, fewer than 1 in 3 contractors report finishing projects on time and within budgets, with 70% citing coordination at the site of construction as the primary reason. In the words of one industry executive, “all [of the nation’s] major projects have cost and schedule issues … the truth is these are very high-risk and difficult projects. Conditions change. It is impossible to estimate it accurately.” But can process improvements and other innovations make construction cheaper, better, faster, and easier?

Example topical areas include but are not limited to: augmented forecasting and modeling techniques; prefabricated or advanced robotic fabrication, modular, and adaptable structures and systems such as bridge sub- and superstructures; real-time quality control and assurance technologies for accelerated construction, materials innovation; new pavement technologies; bioretention; tunneling; underground infrastructure mapping; novel methods for bridge engineering, building information modeling (BIM), coastal, wind, and offshore engineering; stormwater systems; and computational methods in structural engineering, structural sensing, control, and asset management. 

Key Questions for Consideration:

• What innovations are more critical to the accelerated construction requirements of the future? 

Logistics

Our national economic strength and quality of life depend on the safe and efficient movement of goods throughout our nation’s borders and beyond. Logistic systems—the interconnected webs of businesses, workers, infrastructure processes, and practices that underlie the sorting, transportation, and distribution of goods must operate with efficiency and resilience. . When logistics systems are disrupted by events such as public health crises, extreme weather, workforce challenges, or cyberattacks, goods are delayed, costs increase, and Americans’ daily lives are affected. The Biden Administration issued Executive Order 14017 calling for a review of the transportation and logistics industrial base. DOT released the Freight and Logistics Supply Chain Assessment in February 2022, spotlighting a range of actions that DOT envisions to support a resilient 21st-century freight and logistics supply chain for America.

Topical areas include but are not limited to: freight infrastructure, including ports, roads, airports, and railroads; data and research; rules and regulations; coordination across public and private sectors; and supply chain electrification and intersections with resilient infrastructure. 

Key Questions for Consideration:

• How might we design and develop freight infrastructure to maximize efficiency and use of emerging technologies?
• What existing innovations and technologies could be introduced and scaled up at ports to increase the processing of goods and dramatically lower the transaction costs of US freight?
• How can we design systems that optimize for both efficiency and resilience?
• How can we reduce the negative externalities associated with our logistics systems, including congestion, air pollution, noise, GHG emissions, and infrastructure degradation?

To date, Japan’s peaceful nuclear energy use has taken the form of a nuclear fuel recycling policy that reprocesses spent fuel and effectively utilizes the plutonium retrieved in light water reactors (LWRs) and fast reactors (FRs). With the aim to complete recycling domestically, Japan has introduced key technology from abroad and has further developed its own technology and industry. However, Japan presently seems to have issues regarding its recycling policy and plutonium management.

Image courtesy of Shutterstock

Because of recent increasing risks of terrorism and nuclear proliferation in the world, the international community seeks much more secure use of nuclear energy. All of the countries that store plutonium (which can be used for making nuclear weapons) must make the best efforts possible to decrease it. Taking this into account, concerns about Japan’s problem of plutonium management have been growing in the international community and Japan’s accountability for its recycling policy is essential.

In this paper, Yusei Nagata, an FAS Research Fellow from MEXT, Japan, analyzes U.S. experts’ opinions and concerns about Japan’s problem and considers what Japan can (and should) do to solve it.

Read the full report here.

Although the United States still has the largest number of nuclear power plants in the world, it does not dominate global nuclear power. While the United States was the leading nuclear power supplying nation more than thirty years ago—at least for states outside of the Soviet sphere of influence—the reality today is clearly that the U.S. nuclear industry is only one of several major suppliers. The United States can no longer build a large nuclear power plant on its own. Foreign nuclear companies own major U.S. nuclear power companies.

In addition, the United States no longer supplies the majority of the world’s enriched uranium for nuclear fuel; instead, the United States Enrichment Corporation has shut down its enrichment plants based on gaseous diffusion and has been struggling to commercialize the American Centrifuge Project partly due to reduced global demand for enriched uranium and also due to competition from established enrichment companies.

Nonetheless, the United States continues to have great influence on the nuclear market because many of the major supplying nations have built their nuclear power programs on the basis of U.S. technology. In a new issue brief, FAS President Dr. Charles Ferguson takes a look at options for the United States to gain back leadership via a cooperative approach. The brief analyzes what nations could be effective partners for the United States in furthering nonproliferation while providing for the continued use of peaceful nuclear energy. The nuclear industry is increasingly globalized and the United States needs to partner with allies and other nations to advance nonproliferation objectives.

View Full Brief

Global biosecurity engagement programs are designed to prevent the harmful use of biological agents and pathogens. It is difficult to measure the effectiveness of these programs in improving biosecurity given that there have been relatively few attempts to misuse the life sciences. Metrics that focus on outputs (what was done) as opposed to outcomes (the impact of what was done) have not been helpful in determining how these efforts might be improved in the future. As a result, the goals of the programs have traditionally been more quantitative in nature – for example, increasing the number of agents secured and number of scientists engaged. Broadening the scope of biosecurity engagement metrics can help align program goals with a more qualitative approach that prioritizes the international partners’ global health security.

To understand how biosecurity engagement is conducted and evaluated, Michelle Rozo, Ph.D. candidate at Johns Hopkins University, interviewed more than 35 individuals in the United States and abroad (including government officials and their non-governmental partners) regarding current and future programs that can be used to create a cohesive, global health system approach to biosecurity. The results from the interviews are complied in an issue brief which also provides a strategy for policymakers to focus more on qualitative public health outcomes instead of quantitative security outputs. With this strategy, programs will cost less and be more effective in reducing global threats.

View Full Brief

The 2011 accident at the Fukushima Daiichi nuclear power plant was preventable. The Great East Japan earthquake and the tsunami that followed it were unprecedented events in recent history, but they were not altogether unforeseeable. Stronger regulation across the nuclear power industry could have prevented many of the worst outcomes at Fukushima Daiichi and will be needed to prevent future accidents.

In an FAS issue brief, Dr. Charles Ferguson and Mr. Mark Jansson review some of the major problems leading up to the accident and provides an overview of  proposed regulatory reforms, including an overhaul of the nuclear regulatory bureaucracy and specific safety requirements which are being considered for implementation in all nuclear power plants.

View Full Brief

A country with few natural resources, first Japan began to develop nuclear power technologies in 1954. Nuclear energy assisted with Japanese economic development and reconstruction post World War II. However, with the fear of lethal ash and radioactive fallout and the lingering effects from the 2011 accident at Fukushima-Daiichi Nuclear Power Plant, there are many concerns related to Japanese nonproliferation, security and nuclear policy.

In a FAS issue brief, Ms. Kazuko Goto, Research Fellow of the Ministry of Education, Culture, Sports, Science, and Technology of the Government of Japan, writes of Japan’s advancement of nuclear technologies which simultaneously benefits international nonproliferation policies.

Download Full Brief

The nuclear programs of North Korea and Iran have been, for many years, two of the most pressing and intractable security challenges facing the United States and the international community. While frequently lumped together as “rogue states,” the two countries have vastly different social, economic, and political systems, and the history and status of their nuclear and long-range missile programs differ in several critical aspects.

The international responses to Iranian and North Korean proliferation bear many similarities, particularly in the use of economic sanctions as a central tool of policy. Daniel Wertz, Program Officer at the National Committee on North Korea, and Dr. Ali Vaez, former Director of the Iran Project at the Federation of American Scientists, offer a comparative analysis of U.S. policy toward Iran and North Korea in a FAS issue.

Download Full Brief

Charles P. Blair, Senior Fellow on State and Non-State Threats, interviewed Federation of American Scientists’ Senior Fellow for Nuclear Policy Dr. Robert Standish Norris. The report takes a deeper look at the nuclear policies of the Obama administration—polices that Dr. Norris terms “radical” with regard to their vision of a nuclear weapon free world.

Download Full Brief

A modified U.S. nuclear bomb currently under design will have improved military capabilities compared with older weapons and increase the targeting capability of NATO’s nuclear arsenal. The B61-12, the product of a planned 30-year life extension and consolidation of four existing versions of the B61 into one, will be equipped with a new guidance system to increase its accuracy. As a result, the U.S. non-strategic nuclear bombs currently deployed in five European countries will return to Europe as a life-extended version in 2018 with an enhanced capability targets.

Download Full Brief

While diplomats and officials claim Iran has slowed down its nuclear drive, new analysis shows that Iran’s enrichment capacity grew during 2010 and warns against complacency as five world powers resume talks.

Download Full Brief

Since February 2010, Iran has been enriching uranium to concentrations of 20 percent U-235. A stockpile of 130 kg of 20 percent enriched uranium would reduce, by more than half, Iran’s time to develop a bomb. A key unknown is whether Tehran will stop the higher enrichment and, if so, under what circumstances.

Download Full Brief