David Winickoff, OECD Directorate of Science, Technology and Innovation, Division of Science and Technology Policy
Science and Technology Studies scholars have described for some time how global institutions are key sites for the co-production of knowledge, technologies, discourses and authority at a global scale. The adjudication of GMOs at the WTO, the science of environmental impact assessments at the World Bank, the use of self-governance among National Academies of Science, and peer review at the IPCC are just some examples. More recently, we see in bioethics and epidemiology at the World Health Organization, and in the patterns of vaccine production and use, the power of expertise, technology and governance in remaking global order.
To speak of “the” global order is in fact misleading. Perhaps now more than ever, global governance operates in more distributed ways. Governance proliferates through complex networks and nodes. Multiple centers can take on similar and overlapping roles and come into conflict. Traditional alliances are in flux, as states--supported by populist movements--are re-asserting themselves against more internationalist approaches. New and dueling discourses of “technological sovereignty” and “open innovation” mark new tensions in the co-development of natural phenomena, technology and governance.
So, global order is decidedly plural. How might a system of global governance for emerging technology operate, especially given the technological field is moving quickly and given the difficulties of enacting transnational law? If regulatory approaches to global issues, from climate change to food safety, are increasingly operating through pluralistic, overlapping, and fragmented regimes, the challenge of observing and engaging technology governance is the more formidable.
Here I offer a few reflections on the capacities and incapacities of soft law as a mechanism for intervening in this complex landscape. I will do this in one particular context: efforts of the Organization for Economic Cooperation and Development to develop a soft law of technology governance.
The OECD and International Soft Law
Some nodes of global governance are more densely networked than others. Arguably, the Organisation for Economic Cooperation and Development (OECD) is one of these more densely linked nodes. With origins as the administration agency of Marshall Plan, the OECD is a body of 37 countries with advanced economies that share a commitment to “Bretton Woods”-style liberal democratic principles.
The OECD carries out multiple functions which, when combined, distinguish it on the international landscape. First, it operates as a diplomatic space, where countries come together to negotiate sets of common actions such as an agreement on international profit shifting. Second, it is a research institution meant to create a common knowledge base through research and data collection for public policy across countries to help facilitate cooperation and exchange.
Third, it is a forum for making soft law. Its nearly five hundred OECD Recommendations, adopted by the consensus of OECD members, function as normative frameworks for a large array of policies.
Regulation refers to a system of norms that define permissible and impermissible activities with sanction or incentives to ensure compliance. In contrast, soft law refers to policy instruments with moral or political force but without legal enforceability. OECD Recommendations would qualify as soft law. According to the OECD:
OECD Recommendations are not legally binding but practice accords them great moral force as representing the political will of Adherents. There is an expectation that Adherents will do their utmost to fully implement a Recommendation.
Examples of soft law include private standards, general policies, guidelines, principles, codes of conduct, and forums for transnational dialogue.
Developing international soft law at the intersection of science, technology and innovation is one of the OECD’s activities. Over the last few years, the OECD has developed a set of soft law instruments related to the governance of emerging technologies. The OECD Recommendation of the Council on Artificial Intelligence is one example. It helped shape the G7 agenda on AI, and has formed the basis of the new international Global Partnership on AI and the creation of an AI Policy Observatory.
In December 2019, OECD countries adopted a soft law instrument addressing neurotechnology. The OECD Recommendation on Responsible Innovation in Neurotechnology seeks to anticipate problems in the course of innovation and steer technology to better outcomes, while including many stakeholders in the innovation process. The Recommendation is the first international instrument in the neurotechnology field. Over a period of over five years, the OECD led a series of multi-stakeholder workshops that explored strategies for the responsible development and use of brain-computer interfaces and other technologies. It is soft law, non-binding as a legal matter but enforced through moral suasion and regular monitoring across countries.
The Rule of (Soft) Law?
Given the complexities and potentialities of global governance, could soft law make a contribution to forms of good governance, both in substance and process? Are the various instruments of soft law well suited to the governance of emerging technologies where there is often a need to operate across borders and across scales of government, and through networks of power?
Flexibility and Constraint
Soft law could never replace regulation. Soft law carries moral force, underwritten by duties of self-reporting, and provides an ethical basis and incentive for compliance. But countries are able to derogate without legal penalty. The approach is called “soft law” for a reason, as “hard” mechanisms like legal liability do not exist.
However, soft law carries advantages. The various instruments of soft law might be well suited to the governance of emerging technologies where there is often a need to operate at the global scale at a time when the prospect for new international treaties is dim. A more flexible approach might be appropriate given the uncertain trajectories of emerging technologies. Its non-binding nature makes it easier to adopt. Soft law can be pitched at both general and specific levels. A more general approach can allow the development of more directive specificities over time. The modality gains moral force by being able to claim “consensus” with respect to the adopted text.
Soft law can also evolve and harden over time.
Thus, soft law occupies a kind of a central position between flexibility and constraint, international contract and technological sovereignty that presents both strengths and weaknesses in addressing the challenge of global governance of emerging technology.
Legitimacy and Evidence
Democratic legitimacy is a problem for soft law as it is for many other modalities of global governance. “Global publics” such as communities of stakeholders “engaged” by UN processes are necessarily partial and exclusionary. Soft laws such as guidelines and recommendations are often enacted by one authority, yet govern others. Perhaps as a direct consequence of this legitimacy deficit, centers of power shore up their claim by recourse to other forms of authority, especially expert or knowledge-based ones. For instance, in certain areas of international trade law like technical barriers to trade or food safety regulation, the authority of the WTO rests on the work of particular other expert-based international organizations.
At the OECD, legitimacy within the membership resides in the shared ethos of the post-WWII liberal democratic order. But it also resides in the process of joint knowledge production and a shared ethics of evidence. The empirical base, as it is sometimes called, of new soft law must “go through committee”--i.e., to be reviewed by the membership in what is a mix of peer review and political oversight. Mobile forms of quantitative data and its various displays are particularly persuasive and tend to ‘”speak for themselves.” Consensus on the validity and acceptability of the empirical base lays groundwork for the agreement of norms.
There is of course a difference between internal legitimacy of norms within a particular institutional locus, and legitimacy in the broader society. It matters greatly which organization advances soft law. Political trust, shared knowledge and evidentiary practices are grounds for internal legitimacy, but external legitimacy requires more. Legitimacy will be relative to who gets to sit at the table, and who is involved in knowledge production. The OECD Recommendation on Neurotechnology illustrates how the legitimacy of soft law depends on the manner in which it is co-produced with knowledge. That is a lesson from OECD. Legitimacy may require due process in the production of knowledge and expert input.
Normative Layering
The proliferation and overlap of soft law poses a problem. Indeed, because any entity or group can, in theory, develop or propose soft law--whether a political, technical or economic community--there can be a proliferation and even competition of norms. For instance, the EU, the UN, and other bodies are actively developing codes on AI governance. Countries, companies and citizens may have a difficult time in navigating the tangle of soft law programs and traditional regulatory programs. They may also be able to “forum shop” for the soft laws that suit them.
But perhaps multiplicity and overlap could be a strength. Perhaps we don’t want single sources of soft law, operating in a world of unipolar epistemologies. One might see bodies of soft law in any given technology domain as a kind of lattice, where norms can be developed in one context and reinforced in others. The result might accomplish something akin to the generation of international customary law--obligations arising from established international practices rather than treaties.
An Ecosystem of Deliberation
Such an international customary law of technology governance could include many norms of epistemic due process. Deliberation and/or public debate is one such candidate norm. Deliberation goes beyond “public engagement” in that it posits a reasoned discourse for issues of public life. Such a customary international law of tech governance might involve the need to promote cultures of deliberation with regard to emerging technologies.
The OECD neurotechnology recommendation attempts to accomplish this through the duty of members to “Enable societal deliberation on neurotechnology” (Principle 5). Further, in the instrument, OECD Council (its highest governing board) instructs the OECD members that the instrument should:
serve as a forum for exchanging information on neurotechnology policy and experiences with the implementation of this Recommendation and foster multi-stakeholder and interdisciplinary dialogue on innovation in neurotechnology (Article VIII, a).
One might draw together multiple soft law codes to help open diverse but connected spaces that form a kind of ecosystem of deliberation in global governance. The Council of Europe Bioethics Committee has waged a strong project around a cosmopolitan norm of the need to promote Public Debate of New Technology, including a comprehensive guide and catalogue of practices. In October 2005, the General Conference of UNESCO adopted by acclamation the Universal Declaration on Bioethics and Human Rights (UDBHR), which among many other things aims to “foster debate, education and public awareness of, and engagement in bioethics” (Art 19, d).
Such soft law norms can of course reinforce and be reinforced by national hard law. For example, Article 46 of the 2011 French law on bioethics stipulates that the National Consultative Ethics Committee (CCNE) organizes public debates and consultations within the scope of bioethical questions on medical and/or research practices affecting the human body and the embryo. A broad bioethics debate recently occurred in France.
Conclusion: Building Tents and an Institutional Ecosystem
I want to come back to where this essay started, that is the complex and overlapping nature of global governance in the current moment. This structure poses real challenges for enacting legitimate and effective technology governance. I have tried to discuss, in the context of the OECD, the modalities of soft law in this landscape and some of the possible advantages and disadvantages of its ability to guide practice.
There is a persistent danger, of course, that soft law could be seen as a sufficient approach to emerging technology governance, replacing the need for a sound regulatory framework. But to the extent that diverse bodies of soft law can open up spaces of exchange and deliberation, they might be useful in all technology governance contexts, including that of gene editing.
What kinds of institutional and normative ecosystems are necessary to help support a broader and more diverse observation of and engagement in technology governance? Diverse communities of states, sectors and publics must be part of such an ecosystem. Perhaps across bodies of soft law and their communities, there can be interaction, exchange, and mutuality. Such an ecosystem of institutions might build up such an international customary law of technological due process. Here the aim would be less regulatory convergence, than an honest and “reciprocal commentary, mutually deepening."