Nanolessons for Revamping Government Oversight of Technology

In recent decades, the capabilities of U.S. federal agencies responsible for environmental health and safety have steadily eroded. The agencies cannot perform their basic functions now, and they are even less able to cope with the new challenges being created by rapid advances in science and technology. Nanotechnology provides a perfect case for considering ways to redesign the way in which government oversees science and technology in order to protect public health and safety.

Nanotechnology involves working at the scale of single atoms and molecules. The nanoscale is roughly 1 to 100 nanometers. For comparison, there are 25.4 million nanometers in an inch, or 10 million nanometers in a centimeter. Nanoscale materials can have different chemical, physical, electrical, and biological characteristics than their larger-scale counterparts, and they often behave differently than do conventional materials, even when the basic material (say, carbon or silver) is the same.

Their tiny size and novel characteristics offer considerable promise for developing valuable new products. Indeed, many observers suggest that almost every area of human activity will be affected by future nanotechnologies. Medicine, food, clothing, defense, national security, environmental cleanup, energy generation, electronics, computing, and construction are among the leading sectors that will be changed by nanotechnology innovations.

On the flip side, however, the novel characteristics of nanomaterials mean that risk assessments developed for ordinary materials may be of limited use in determining the health and environmental risks of the products of nanotechnology. Although there are no documented cases of harm attributable specifically to a nanomaterial, a growing body of evidence points to the potential for unusual health and environmental risks. This is not surprising. Nanometer-scale particles can get to places in the environment and the human body that are inaccessible to larger particles; and as a consequence, unusual and unexpected exposures can occur. Nanomaterials also have a much larger ratio of surface area to mass than do ordinary materials. It is at the surface of materials that biological and chemical reactions take place, and so it is to be expected that nanomaterials will be more reactive than bulk materials. Such new exposure routes and increased reactivity can be useful attributes, but they also carry the potential for health and environmental risk.

Starting from scratch

The government’s current system for overseeing technologies and their spinoffs was designed to deal with the problems of steam engines in a precomputer economy-that is, yesterday’s world. It was based on assumptions that most problems are local, that programs can be segmented and isolated from each other, and that technology changes slowly. All of these concepts are no longer valid, if they ever were.

Given this woeful state, combined with the diverse challenges that nanotechnology and other rapidly advancing fields often pose, it would be a mistake to imagine that simply tinkering with the existing oversight system will suffice. Only a complete overhaul will succeed. The government needs to adopt a new organization that will operate under new legal authority and be equipped-and funded-to use new tools that will provide the knowledge and flexibility required for effective oversight. At the same time, however, new oversight requirements should be applied with a constant awareness of the need to encourage technological innovation and economic growth.

As a first step, the government should create a new agency, which might be called the Department of Environmental and Consumer Protection, in light of its expanded mission. It would incorporate six existing agencies: the Environmental Protection Agency (EPA), the U.S. Geological Survey (USGS), the National Oceanic and Atmospheric Administration (NOAA), the Occupational Safety and Health Administration, the National Institute of Occupational Safety and Health, and the Consumer Products Safety Commission. It would also incorporate a number of new units devoted to particular tasks that now get short shrift or are scattered across agencies. Many of the components would be allowed to operate with a good deal of independence. The success of the new organization would depend greatly on the degree to which it could strike a good balance between the integration and independence of the components.

The agency would differ from existing agencies in that it would be a science agency with a strong regulatory component rather than a regulatory agency with a science component. The emphasis on science is necessary to deal with the rapidly changing technologies that require oversight.

The agency would be among the smaller federal cabinet departments, but not the smallest. It would employ approximately 43,500 full-time equivalent personnel and have an annual budget of roughly $18 billion. This would make it 10 times larger than the Department of Education and 4 times larger than the Department of Housing and Urban Development. However, it would be half the size of the Treasury Department and a quarter the size of the Department of Homeland Security.

Importantly, the agency would be significantly larger than the EPA or any of the other federal oversight agencies. One advantage of housing oversight functions in a larger organization is that it increases overall flexibility. In the smaller agencies, resources are devoted primarily to ensuring their survival and to the performance of the minimal required functions; they have limited ability to anticipate and respond to new problems or to consider new ways of doing things. Also not to be overlooked, the current small size of the regulatory agencies makes them vulnerable to becoming even smaller. “The large get larger” seems to be the organizational analog to the rich getting richer.

Within this broad framework, the agency would focus oversight on products and pollution, and do so in a more integrated way. For example, its activities would include:

Product regulation. The current government oversight system tries to focus on both materials and products, often varying by agency, but this hodge-podge serves neither goal very well. Materials can be thought of as raw ingredients with their own sets of properties. But the way in which a material is used or how it is combined with other materials often determines whether adverse effects will occur. Therefore, focusing on materials alone will not provide a sufficient basis for evaluating risk. Products, on the other hand, are the items made from the materials and sold to public consumers, manufacturers, or others. A product may go through multiple stages, with each stage being a separate product. For example, carbon nanotubes (one product) can be combined with plastic in a compound used for car bodies (a second product), and that compound can be used in a finished automobile (a third product). Because the specific characteristics of specific products are likely to determine the adverse effects that might occur, future oversight will need to focus primarily on products.

Such a product-focused system must incorporate at least two principles. First, oversight should encompass the life cycle of the product through manufacture, use, and disposal. (Transportation is also part of the life cycle, but it can be regulated separately by the Department of Transportation.) Second, the degree of oversight-that is, the stringency of regulatory requirements-should be related to the anticipated harm the product will cause. This is a function of the severity of anticipated harm and the likelihood that it will occur. The government is not likely to have detailed and current information about the composition of a product, its intended use, or its anticipated effects. Only the manufacturer will be able to know or obtain this information on a timely basis. Thus, the government inevitably must depend on the manufacturer to test the product and accurately report relevant information. The government should put in place strong penalties for distorting, concealing, or failing to obtain required data, in order to deter such behavior.

One approach to engaging manufacturers would be to require them to develop a sustainability plan (SP) for each of their products. The plan would contain a summary of known information about the components of the product, the adverse effects of the product, a life-cycle analysis of the product describing its use and manner of disposal, and an explanation of why the product would not cause any undue risk. The government would define as precisely as possible what data are required and what constitutes undue risk, and it could require additional information for particular categories of products.

It seems reasonable to require every manufacturer to know this information before selling its product. (Small businesses should not be exempted, though special efforts will be needed to inform them about the requirements and to provide them with technical assistance to help them meet the requirements.) Moreover, manufacturers would have to update an SP after a product is approved and marketed if they became aware of new information that affected the product’s risk. A number of firms have voluntarily produced statements similar to an SP. For example, DuPont, in cooperation with Environmental Defense, developed a framework for analyzing the risks of nanomaterials and now applies the framework to all of its new nanoproducts.

Because every product, except for a few that would be exempted, would have to have an SP, manufacturers would be able to know the potential risks of components they use by requiring their suppliers to provide them with the SPs for the components. This would be a major benefit to manufacturers of complex products such as automobiles. At present, manufacturers may be legally liable for problems caused by components they use, but they may have no practical way to find out what the risks of the components are.

How such SPs would be used and what additional information, if any, might be required would depend on the harm the product might cause. A possible typology is as follows:

Category 1: This category would cover products that have a low probability of having adverse effects. There would be no oversight; the manufacturer would simply retain the SP, or, if there were clearly no significant risks, the manufacturer might be exempted from the SP requirement altogether. Most products are likely to fall in this category. There is always the possibility, however, that new evidence will move a category 1 product to a different category.

Category 2: This category would cover products for which risk-communication measures should be sufficient to avoid adverse effects. The manufacturer would be required to use the SP as the basis for producing a product safety data sheet to be given to users or for adding labels to products sold to consumers.

Category 3: This category would cover category 1 or 2 products that have been marketed but later become suspected of causing adverse effects. The government would be empowered to halt manufacture or distribution of the product pending a review of its safety.

Category 4: This category would cover products that have some probability of causing adverse health or environmental effects and would therefore be subject to premarket review. Products in this category might include pesticides, fuel additives, and products containing designated types of materials such as persistent organic pollutants.

The government would define the categories and decide which products belong in which categories. To the extent possible, the government would assign broad classes of products to particular categories. If a manufacturer wanted to produce a product that was not included in one of the classes, it would have to submit a request to the government to designate which category the product belonged in. For categories 3 and 4, the burden of proof would be on the manufacturer to demonstrate that the data in the SP were valid and adequate, and that they supported the conclusion that the product would not or did not pose undue risk. The government might have to show some cause for categorizing a product as category 3.

The major challenge in regulating products is the enormous number of products on the market at any given time. For example, the Consumer Products Safety Commission oversees 15,000 types of products, and each type contains numerous individual products. Inevitably, the number of products placed in each category would, to some extent, be determined by the resources available to the government oversight agency. The first two categories would require only spot checking by government, and category 3 probably would apply to only a relatively small number of products. Category 4 would require intensive use of government resources.

To help pay for this system, the government should consider charging manufacturers a fee for seeking product approval, as the Food and Drug Administration now does for drug registration (although steps would need to be taken to avoid some of the problems with that system). Consideration also should be given to making public on a regular and timely basis whatever gap may exist between resources and oversight requirements. This could be done by requiring the agency to regularly publish the number of products that should be reviewed but for which resources were not available to do the review.

Pollution control. Control of nanoparticles released during manufacture must be based on preventing the releases from occurring. Trying to deal with the problem by separately regulating releases to the air, water, or land, as current law does, will not work. Instead, the goal should be to foster the adoption of integrated pollution prevention and control methods that capture or recycle pollutants. In Europe, such technology is facilitated by having a single environmental permit for each facility. In the United States, however, various federal (and often state and local) programs regulate different kinds of pollution and require their own permits. The system not only results in bureaucratic duplication and confusion but also makes the permitting process opaque to the public. Moreover, because of the fragmentation, the system often fails to control a significant portion of a facility’s environmental impact.

Monitoring. Monitoring provides the link between government actions and the real world. Various components of the proposed agency would do two types of monitoring: environmental and human. A number of federal agencies now conduct a range of environmental monitoring activities. Recently, a group of science policy experts proposed combining the two largest monitoring agencies-the USGS and NOAA-into a single independent Earth Systems Science Agency. This idea would be incorporated in the proposed agency, though the new component would be granted a semi-independent status to ensure maximum flexibility. The agency also would take over the EPA’s monitoring functions and would add a new Bureau of Environmental Statistics, analogous to the Bureau of Labor Statistics.

Monitoring human health will be important to spot any products that turn out to cause adverse health effects that were not identified before the products were marketed. Given the uncertainties of risk assessment for new technologies, this situation is almost certain to occur. Such consequences will not be identified unless there is an extensive surveillance system that spots abnormal health phenomena, such as an excess number of cases of a given disease or a spike in emergency room admissions. The system should be coordinated with other domestic and international health reporting systems and it should be as unobtrusive as possible.

Technology assessment. With only a few exceptions such as nuclear power, technology as such is not and should not be regulated in the same sense that products and wastes should be regulated. But it remains critical to fully understand the potential consequences of any technology, in order to guide its application along beneficial lines. The proposed agency can play an important role in conducting and disseminating thorough and balanced assessments that are intended to inform, not promote, and to engage a broad swath of the population.

The need for such assessments may be especially great given the possible social and moral implications of nanotechnology. For example, if it proves possible that nanotechnology can be used to improve the functioning of the human brain, should it be used that way? And if so, for whose brains? If nanoscale materials are incorporated in foods to improve nutrition, taste, or shelf life, should the food have to be labeled to show that nanotechnology has been used? If synthetic biology, using nanotechniques, can create new life forms, should it be allowed to do so?

Focusing on particular applications may miss the overall effects of a technology, and by the time the implications of the applications become clear, it may be too late to effectively influence the direction the technology takes. What is needed is a capability to consider the overall effects of major new technologies and to do so while there is still time to deal with them. This requires a forecasting capability as well as an assessment capability. The techniques for doing forecasting and assessment have not received the attention they need. Not coincidentally, the institutions for making forecasts and conducting assessments are weak or nonexistent.

Moving to a radically different oversight system will require new approaches in a number of other areas as well. These areas include:

Risk assessment. Forecasting the risk of any technology involves basic scientific information about the technology, test data on specific products, and risk assessment. Each of these components has a different source and different characteristics. Basic scientific information comes primarily from university and government laboratories. The motives for developing the information include scientific curiosity, the possibility of obtaining grants and contracts, and the possibility of making money through patents or startup companies. Meeting societal needs, such as identifying the risks of new technologies, is often not a major consideration in setting the basic science agenda. This is one reason why it is important for government oversight agencies to have their own scientific resources.

Testing of specific products is done primarily by their manufacturers, either in house or through contract laboratories. It is beyond the resources of government agencies to test the multitude of products and, in any case, the manufacturer will be most knowledgeable about the products it is making. Testing for new kinds of products can be problematic. For example, it is often not known what end points (cancer, asthma, fish mortality, and so on) to look for when testing nanomaterials, nor is it understood which characteristics of the material are associated with adverse effects. In the absence of testing, conclusions about the safety of a product or material are often based on analogous items that have been tested. However, by definition new types of products do not have exact analogs that have been tested. When technologies are evolutionary, as many nanotechnologies are, analogs may help predict behavior, but they are still generally not an alternative to testing.

The technology of testing is itself changing, and there has been progress in developing tests that are much faster and cheaper than current tests that rely on laboratory animals. The type of risk assessment usually done by the government has evolved into a highly sophisticated set of procedures that are intended to help decisionmakers make rational decisions. But risk assessments typically are not scientific products; they are a way of organizing and analyzing data about a particular substance or product. They are not scientific, because only in unusual cases can they be empirically verified. The typical risk assessment may result in a finding that substance X will produce Y number of additional cancer cases per million people exposed. However, whether Y is 0 or 1,000 in reality will never be known because there are too many other causes of cancer. Regulatory decisions almost always must be made based on the weight of the available evidence. Conclusive scientific proof is usually not to be had, although the better the available science, the easier it is to do a risk assessment and the more accurate the assessment is likely to be.

Because decisions typically must be based on balancing the available evidence, the default assumption about who has the burden of proof is critically important. In the United States, the Toxic Substances Control Act puts the burden on the government to prove the risk. In Europe, however, the burden falls on manufacturers. Industry occasionally argues that the burden should be on the government because it is not possible to prove safety, but this is a fallacious argument. It is not possible to conclusively prove the safety of a product, just as it is usually impossible to conclusively prove the risk. Risk and safety are both operationally defined by required tests, and it is equally difficult to prove either one.

Enforcement. Enforcement has two related dimensions: incentives and compliance. The stronger the incentives, the better the compliance, but the two dimensions involve different considerations. The increasingly rapid pace of technological innovation and the diversity of the innovations have made it difficult to apply many of the older enforcement approaches. Newer approaches have emphasized economic incentives and flexibility. Liability has been used as the major incentive in one U.S. waste law (the Comprehensive Environmental Response, Compensation, and Liability Act of 1980), and it might be possible, for example, to make manufacturers legally liable for failure to develop an SP or for any adverse consequences that could reasonably have been foreseen but were not included in the plan.

A downside to using liability and litigation in implementing regulatory oversight is that government employees might have to spend large amounts of time giving testimony in court, making depositions, and participating in litigation in other ways. This might seriously affect their ability to perform their primary duties. Cap-and-trade programs, such as the one the federal government uses to regulate sulfur dioxide emissions from power plants, have been proposed as a substitute for much of the existing pollution control structure. Effluent fees and charges also have been used in a few situations and have been suggested as an approach that could be used more widely. Whether these kinds of approaches could be used for oversight of useful products (as contrasted with wastes) is not clear, and at the least, caution must be exercised when proposing that incentives developed for curbing wastes be applied to useful products.

Insurance is another incentive that can be important. It can be used either negatively or positively. Negatively, one insurance company has already refused to insure for any damage connected with nanotechnology, citing the lack of adequate risk information. If other companies follow suit, this could be a major incentive for more research and more testing of products by private firms. Insurers could deny insurance to manufacturers that do not have an SP. On the positive side, insurance could be given to manufacturers against tort suits if the manufacturer had an adequate SP and had implemented that plan, and the tort suit covered a subject that was included in the plan.

With respect to compliance, the key question probably is the extent to which one can rely on voluntary compliance. The answer depends on the cultural context. In the United States, oversight in many contexts has shown voluntary compliance to be undependable. Legally enforceable requirements, vigorously implemented, are necessary to deal with the usually small, but important, percentage of firms that are not good corporate citizens.

International cooperation. The combination of a worldwide economy and near-instantaneous global communication has made technology oversight an international issue. Every oversight function, from research to enforcement, now has important international dimensions. The challenge is how to embody the international dimensions in effective institutions. The European Union oversees many issues related to technology within its member nations. The Organization for Economic Cooperation and Development (OECD), which includes most of the industrialized nations, has taken a variety of initiatives related to new technology. It has agreed to test 14 generic nanomaterials for health and environmental effects, and it has established a database for sharing research information on potential adverse effects of manufactured nanomaterials. The United Nations (UN) also has several components relevant to oversight, including the World Health Organization (WHO), the UN Environment Programme (UNEP), and the International Labor Organization. In addition, many nongovernmental international organizations, including international trade associations and mixed public/private organizations, such as the International Organization for Standardization, play a part in oversight efforts.

In the long run, an international regime for product oversight may develop to match the international trade in products. At the least, U.S. and European regulatory approaches should be made consistent. In the interim, the emphasis should be on information sharing. At least three types of information should be made available internationally: research results on adverse effects of a technology; standards, regulations, and other oversight policies, as well as decisions applied to a product or technology; and reports of any adverse health or environmental effects that occur and could be attributed to a product. The OECD has made a start on the first two. The third is an important function that needs to be supported, perhaps by a joint effort of WHO and UNEP. An international system for reporting adverse effects would have to draw heavily on existing surveillance systems. The current worldwide economic crisis and the collapse of the Doha Round of international trade talks have made the future of all international efforts uncertain. One outcome of the current crisis could be a stronger set of international institutions, even perhaps including the basis for an internationalized system for dealing with new technologies and products.

Public involvement. Transparency should be the hallmark of oversight activities. Without it, the public interest tends to be submerged beneath the interests of bureaucrats, politicians, and special interests. Transparency becomes even more important in the context of new technologies, because if the public senses that secrets are being kept and motives are being hidden, it may reject a new technology regardless of its benefits. As the International Risk Governance Council has noted, new technologies will require more public involvement because their “social, economic and political consequences are expected to be more transformative.” The challenge, as expressed by the UK’s Royal Commission on Environmental Pollution, “is to find the means through which civil society can engage with the social, political, and ethical dimensions of science-based technologies, and democratize their ‘license to operate’… a challenge of moving beyond the governance of risk to the governance of innovation.”

In the United States, the 21st Century Nanotechnology Research and Development Act, the law governing domestic nanotechnology research, requires the National Nanotechnology Coordination Office to provide “for public input and outreach to be integrated into the Program by the convening of regular and ongoing public discussions, through mechanisms such as citizens’ panels, consensus conferences, and educational events, as appropriate.” The National Science Foundation has experimented with some of these techniques, but overall, little effort has gone into implementing this part of the law.

In the context of new technology oversight, the public can be thought of as three groups: the insiders (such as industry representatives, nongovernmental organizations, academic experts, and labor union representatives), the somewhat informed general public, and the bystanders. The majority of the population falls in the category of bystanders. They do not know about or understand most new technologies, and they do not follow what the government does or says about them. However, even the bystanders may influence oversight through their role as consumers, and the products they buy may be influenced by the opinions of the insiders. A goal of U.S. public policy has been to move people from the bystander category to the informed category. This is consistent with a Jeffersonian view of democracy and is an important way of reducing the chances that the public will react against a technology based on propaganda or misinformation. How successful efforts to inform the public can be, what methods can be used, and how to draw the line between information efforts and propaganda remain important areas for study.

Government impact. How the government should influence the direction of new technology is a knotty question. The government exerts a major influence now through financial support for private R&D, appropriations for defense and other science-intensive government programs, and regulations (or the absence of regulations) on various activities. All of these actions are usually taken piecemeal, without any coherent strategy for the overall technological future of the world or even for the future of any particular technology. Consideration should be given to using “social impact statements” analogous to the environmental impact statements required of government projects. The statements would provide a vehicle for the public to learn about new technologies and for both the public and the government to consider what steps, if any, should be taken to maximize the beneficial impact of the technology and to minimize its adverse effects. Who would prepare the statements, when would they be prepared, what would be their scope and level of detail, and how they would be disseminated are all questions that need to be answered.

In addition, individual government agencies need to become more aware of their impact on technological development and of the impact of technologies on society. The foremost example is the military, which has pioneered the development of a large number of significant technologies, ranging from the pesticide DDT to the Internet. The Department of Defense should establish a Defense Technology Review Board to weigh the civilian as well as the military consequences of new military technology. Board members would have to be privy to all aspects of defense R&D. The board would provide advice to the military departments and to the president’s science advisor.

A remodeled oversight program would improve the government’s ability to handle almost all major environmental and consumer programs. For example, it would allow climate change research and modeling to be brought together under one agency under the research and monitoring functions. The same agency would be responsible for controlling greenhouse gases under the oversight function, and the head of the agency could formulate overall climate policy with the benefit of advice from the scientific and regulatory components of the agency.

It is also clear, however, that moving to such a program will not be easy. The political system operates incrementally except when faced with a crisis, and it is to be fervently hoped that no crisis arises with respect to nano- or any other technology. But over the long run, the political system also responds to models of what could or should exist. Goals and ideals, even if a sharp departure from the status quo, can influence the thinking of policymakers and the public. Many of the needed changes will take a decade or more to accomplish, but there is an urgent need to start thinking about them now.