It was my great delight to be able to present a paper, The Algorithmic Law of Business and Human Rights: Constructing a Private Transnational Law of Ratings, Social Credit, and Accountability Measures authored together with Matthew McQuilla (Penn State SIA MIA 2021) at the Algorithmic Law and Society Symposium 2 December 2021.
The paper suggests the consequences of the iconoclasm of emerging principles of algorithmic governance--the move from law as an exogenous force to a system of rationalized trolling realized through the application of dense webs of quantifiable measures. The focus of law shifts in an algorithmic environment in the sense that the social control at the heart of the enterprise of law is now delegated out and its language transformed. Law produces the granular application of normative principles that are give effect to socio-political taboos and behavior expectations. These are realized not through the traditional bureaucratic-administrative apparatus, but rather through its conversion to measurable indicators against which collective and individual behaviors are judged, punished and rewarded. The ratings system environment that increasingly structures the consequences of human agency provides a very nice example of the way in which algorithmic governance both expresses law and managed behaviors directly.
But the emerging apparatus of algorithmic law shares common socio-legal characteristics with the apparatus of traditional law--like legal systems it is supplementing or displacing algorithmic governance appears to converge around three important characteristics.
The first is that is is marked by an increasing division of labor. Those who harvest data may not be the same as those who develop analytics. Different still may be those collectives that produce algorithms--the judgments and interpretations of data and analytics for developing appropriate blacklists, whitelists and rewards and punishments. Specialization has produced division of labor that makes accountability more difficult to coordinate.
Second, like the business of law, algorithmic governance has become a business--and a business that enhances the business models of non-governmental and non-profit collectives. But as a business algorithmic governance also exhibits the core characteristics of market behaviors. These include privileging principles of property and property rights that make transparency more difficult. It also enhances normative fracture as competition among data collectors, data analytics, and providers of interpreters of data and those who craft systems of effective compliance sharpens in the search for influence, and funding.
Third, even as markets for both the production and consumption of ratings as a form of algorithmic governance grow, so does the consolidation of producer groups and perhaps ultimately their control of entry into markets for the production and interpretation of ratings. And indeed one already sees in the elaboration of systems of ratings studied the development of tight interconnections among a small group of financing institutions (foundations, states), data harvesters and analytics providers, and government all of whom contribute and control (1) the goods necessary for the production of ratings and (2) control the parameters for production and valuation.
The PPT of the presentation follows below (and may be downloaded HERE). The draft (including tables and figures of interconnected ratings maker relationships) may be accessed here (the Abstract and Introduction is included below). Matthew and I welcome comments, engagements and suggestions!
The Algorithmic Law of Business and Human Rights: Constructing a Private Transnational Law of Ratings, Social Credit, and Accountability Measures
Larry Catá Backer [1]
Matthew B. McQuilla[2]
(1 December 2021)
Abstract: This paper examines the rise of algorithmic systems—that is systems of data driven governance (and social credit type) systems in the context of business and human rights and its ramifications (especially its challenges) for law. The context centers on the examination of the landscape of such algorithmic private legal systems as it has developed to date by considering the extent to which a rating or algorithmic system has been emerging around recent national efforts to combat human trafficking through so-called Modern Slavery and Supply Chain Due Diligence legal regime and international norms. Section 2 provides a brief contextualization of the problems and challenges of managing compliance with emerging law and norms against forced labor and in its most extreme forms, modern slavery. Section 3 (From Law to Rating Systems--The Theory and Practice of Forced Labor Ratings) examines the landscape of such algorithmic private legal systems as it has developed to date in the context of forced labor ratings systems. There is a focus on the connection between the power to impose the normative basis of data analytics and the increasingly tightly woven in-connection between principal actors in this endeavor. Section 4 then briefly considered ramifications for liberal democratic orders and the constitution of law. Among the more relevant are those tied to issues of privacy, of the integrity of data, and on transparency. .
Type of paper: Conceptual
Main Contribution to Literature: (1) AI, democracy, development, and the rule of law; (2) Algorithmic decision systems.
Key Methodological Considerations: comparison-based research drawing inferences from a review of current practice in light of conceptual framework developed.
1. Introduction; The Temptations of Algorithmic Law
From well before the start of this century, those committed to the development of legal regimes for the regulation of the human rights effects of economic conduct, especially economic conduct across borders, have been frustrated in those efforts in a number of ways (e.g., Ruggie 2013; Weissbrodt 2014). The frustration arises from the misalignments between the way structures of law and of business have developed. The misalignments exist both as a consequence of the jurisdictional limitations of the two systems of management (one based on territory, the other on function (Backer 2012)), and on the language and methods of management (one based on the exogenous proscriptive structures of public law (Backer 2012a), and the other on the endogenous application of metrics based systems of rewards and punishments for values based clusters of activities (Backer 2018)). The resulting policy and regulatory misalignments have created one of the great challenges for any global human rights (and now sustainability) project focused on the coordinated management of economic activity (Aguirre 2004).
At the core of the challenge are the great contemporary battles around the parameters and functioning of the state system that emerged in its current form after 1945 (Henkin 1995). At its center is a contest over the leading role of states as the apex source of (traditional) lawmaking in a global order framed around public international institutions in which states assume a supporting role (Peters 2005). That notion of states at the apex of political authority and legitimacy has been built into the deepest structures of the UN system and has served as the core principle on which the legalization of economic regimes have been built. Those economic regimes, in turn have elaborated the core principles of free movements of goods, capital, and enterprises, with a limited but robust managed movement of labor (Robe 2012; Ruggie 2013). At the same time, the frustrations of governance gaps and misalignments have accelerated two trends. The first is the privatization of the state and the second is the governmentalization of the private sector (Backer 2021). Both public privatization and private governmentalization have been enhanced by the elaboration of compliance oriented data driven systems, based either on a public or private administrative model, now challenge the role of traditional legal systems as the primary mechanics for the description and enforcement of conduct rules and expectations, and in the process appear to be contributing to the role of traditional legal mechanism in compliance dominated legal cultures (American Law Institute 2020).
These frustrations and challenges, then, might usefully be organized in nine broad categories. Taken together, these suggest the possibility of a potentially radical transformation of governance. That transformation is made necessary by the constraints now built into the construction of the international order, by the limitations of law to reach and govern certain activities, by the substantial normative constraints represented by corporate law principles, by the movement toward compliance based culture, and by the possibilities technology now offers for real time reward-punishment systems built on data-driven systems through which machines can apply analytics based algorithms that produce ratings on which immediate punishments and rewards may be applied (Backer 2018).
First, the normative principles on which economic behavior is conceived, and now reified in the corporate (or enterprise) form, itself have served to limit the way in which the human rights effects of economic activity are embedded and valued within the process of production. The principle that responsible business conduct, or corporate social responsibility, is an externality—that is that it lies outside of the way in which the core of business activity is operated, and its costs measured—remains a challenge for those who would seek to develop mechanisms for internalizing the costs of such conduct (Shum and Yam 2011). The notion of business activity as limited to the production of profit for shareholders in strict compliance with law (Friedman 1970) remains strong even as the nature of the process of production and the role of business enterprises in social and economic collectives changes (Robé 2012) and with it the turn to the fundamental ordering principles of stakeholder governance and a broader understanding of compliance (e.g., Interligi 2010; Munilla and Miles 2005).
Businesses play a vital role in the economy by creating jobs, fostering innovation and providing essential goods and services. Businesses make and sell consumer products; manufacture equipment and vehicles; support the national defense; grow and produce food; provide health care; generate and deliver energy; and offer financial, communications and other services that underpin economic growth (Business Roundtable 2019).
The stakeholder model is itself now more at the center of discussion, even if the primary purpose of the enterprise is to deliver value to its investors. Increasingly that value must be understood within the context in which the enterprise operates, and taking into account its effects on those touched by the processes of production (Crilly, Ni and Jiang 2016).
Second, even if one can overcome or transform core corporate principles (Global Compact Network Germany and Twentyfifty Ltd 2014), traditional legal systems are constrained by the limits of territory and the ideologies of territoriality (Zerk 2010). Jennifer Zerk noted that states tend to resist measures that constrain their ability to manage their own macro-economic policies through law or otherwise; that convergence can tend to expose the inequality between developed and developing states in their ability to contribute to the construction of norms and methods; that multilateral extraterritorial application can be enhanced and its legitimacy enhanced through robust processes of engagement; and that a tort based (compensatory) model holds promise (Ibid., pp. 201-203). At the same time she noted, and quite correctly, that states prefer extraterritorial application of their own norms along control chains (in the context of the activities of sovereign wealth funds see, e.g., Backer 2013), and for multilateral cooperation to manage cross border risk (Zerk 2010, 204). Academics have considered the extent to which extraterritoriality is inevitable under international law in the business and human rights context (McCorquodale and Simmons 2007; but see Van Schaack 2014). They have developed principles to manage its application in this context (Maastricht Principles on Extraterritorial Obligations of States; discussed in Salomon 2012). They have explored the extent to which extraterritoriality—especially when practiced by powerful developed states with the “right” normative baseline, can or ought to serve as the “magic potion” to enhance corporate accountability (Bernaz 2013). Extraterritoriality is viewed as a means of avoiding the impunity that arises from the governance gaps that are the inevitable consequence of global production chains overlying state based territorially constrained authority (Deva 2004).
Third, consequential governance gaps have produced strong soft law measures (Backer 2008). While there was some effort, after the U.S. Supreme Court decision in Kiobel (2013, discussed in De Schutter, et al. 2012; Salomon 2012) to shift the focus of transforming the United States from a national to a transnational court with global jurisdiction for the litigation of human rights claims against enterprises (Fairhurst and Thomas 2015), some attention was also shifting to the strengthening of soft law international regimes. These, some felt, might then be hardened by requiring that they be incorporated into the private law internal governance regimes of large multinational enterprises (Ruggie 2008, ¶¶ 51-81).
Fourth, states have been reluctant to fill the governance gaps required to align global production and global legal measures through the construction of harder multilateral legal measures. The history of movements toward the internationalization, in law, of corporate governance, has been one marked by substantial failures. The near generation long progress from the creation of codes of conducts for multinational enterprises (Chance 1978), to abandonment of the effort to adopt a United Nations Code of Conduct on Transnational Corporations through the 1980s (discussed in Savant 2015), and to the rejection of the Norms on the Responsibilities of Transnational Corporations and other Business Enterprises (Backer 2006; Weissbrodt and Kruger 2003) provides its history.
Fifth, contemporary efforts to develop traditional legal frameworks through international law making have now run up against emerging fracture in the approaches to law (Amstutz 2008). The fracture is not merely between public and private, state, international, and transnational legal “spheres” (Backer 2012a); this to some extent represents a new expression of an older form of allocating political and regulatory authority (Grossi 2004, pp. 23-54). But more importantly, it is one in which Western, Marxist Leninist and developing state conceptions of law and its objectives no longer align (Backer 2017).
Sixth, the ideology of traditional law, which centers the state, the process-based enactment of law and public enforcement though an administrative and judicial apparatus is becoming increasingly detached from discretion-based regulation of public administrative organs. These range from the management of behavior through compliance letters that seek to manage the exercise of administrative discretion on compliance measures (the US Justice Department letters on the exercise of discretion in prosecution (Backer 2019)), to the development of measures against which to assess compliance with administrative objectives, most controversially, the use of the “dear administrator” letters by officials of the Obama Administration Department of Education on university approaches to allegations of sexual assault on university campuses (Brown 2017). States, like enterprises, are increasingly relying on data driven accountability structures to enforce standards. However, in the process the actual interpretation and gap filling functions, that is the movement from legal objective to the actual standards against which conduct is judged, has shifted from the text of law to the analytics of data.
Seventh, states have increasingly sought to use markets as a means of extending compliance based regulatory governance. In the human rights field these include the adoption of disclosure regimes (generally Backer 2008). Most recently, disclosure and disclosure regimes have focused on specific human rights related violations—from human trafficking (Modern Slavery Act [UK)] 2015, Part 6 ¶ 54 [Transparency in Supply Chain]; Modern Slavery Act 2018 [Aus.]), to violations in the mining sector (Dodd Frank Act, §1502). This approach has been criticized by those who view markets with suspicion as a regulatory vehicle (e.g., Garcia 1999). Yet in its modern form it seeks to use the market as a disciplinary mechanics operationalized by data driven judgments grounded in reporting; more importantly it has been viewed as a first step in the mandatory imposition of human rights obligations on enterprises by states in a more comprehensive way (Redmond 2018). These have sometimes been subject to state constitutional limitations, especially where enterprises are compelled to engage in commercial speech (National Association of Manufacturers 2015; Taylor 2017). Despite the controversy, the adoption of disclosure based approaches has opened the door to administration of conduct, not through command (which is now transformed into objective or risk avoidance parameters), but as a function of judgment made possible by data analytics assessed against governing norms.
Eighth, the development of transnational private law based managerial structures for enterprise self-regulation based on compliance and risk prevention, mitigation and remediation strategies, has become an increasingly important element of corporate internal regulatory culture, and one that aligns with the disclosure based legal obligations imposed by states (American Law Institute 2019).[3] These build on two principal developments. The first is the increasing detachment of law from governance, and government from the state (Backer 2012). The second is grounded in the increasing devolution of regulatory power to administrative agencies, on the one hand, and to private actors, on the others. But this is a devolution of a regulatory character, one in which law or norms mandate objectives and principles, but in which the administrative agency or (increasingly the private entity) has responsibility for developing regulatory systems within their production chains for which they are now accountable to state or international organs (Backer 2008). Over the last decade, these, in turn, are based on the recognition of measurable factors from which compliance assessments can be made (Backer 2018).
Ninth, in Marxist Leninist regimes, law itself has been challenged as the principal means of ordering behavior in favor of data driven assessments of conduct producing systems of rewards and punishments based on data analysis tied to measurable social objectives. The leading example, the Chinese Social Credit system, is based on a series of guiding opinions developed and managed through the Chinese State Council especially since 2014 (China State Council 2014), with the objective of building a comprehensive state-based system of law (understood as the guidance of behavior grounded in social, moral, and political principles and correction of unacceptable behavior deviations (Backer 2018). These social credit systems now extend to enterprises within China and may be projected outward through China’s Belt and Road Initiative (Huang 2016; Chatzky and McBride 2020) partner states (Mardell 2019; Brattberg and Soula 2018). The result has been a movement toward ratings based private systems of compliance and accountability in the West, and toward the imposition of public ratings based social credit compliance systems in China.
Thus framed, the consequences for the organization and management of human collectives are potentially significant. At its most fundamental, these trajectories now appear to shift the character of law and lawmaking to one centered on the constitution and operation of accountability structures around public agencies and private actors responsible for building and operating ratings producing systems. If lawmaking were reduced to compliance structures and normative principles/objectives, a quasi-constitutional (certainly constituting) role, it might then follow that the actual task of making rules and applying them will either fall to administrative agencies or private actors exercising delegated power or complying with public (disclosure and operational) mandates (Krajewski, Tonstad, and Wohltmann 2021). In either case, those public or private officials will exercise effective regulatory power of unspecified dimension articulated through the decisions made with respect to data, analytics and ratings. Law, then, assumes a substantially distinctly constituting and auditing character. Traditional law shifts, a least in this context, from the source of rules to the means through which algorithmic law structures are constituted and its jurisdiction and objectives defined (a para-constitutional role). Traditional law becomes the basis for operating operational quality control systems (e.g. Evans and Lindsay 2002) on targeted aspects of societal activities. Algorithmic law is what follows from the constitution of data driven implementation, accountability, or compliance systems. It expresses law through the judgment derived from the analysis of data that is itself driven by the valuation inherent in the judgment expressed through law. Law provides the objective and the normative principle; it is macro-law; algorithmic law translates that into the minutiae of everyday actions—it serves as micro-law. But it is micro law with a language of its own.
Building on those insights, this paper considers the use of algorithmic systems—that is systems of data driven governance (and social credit type) systems-- in the context of the regulation of the human rights effects of economic activity and its ramifications (especially its challenges) for law. We start with a brief examination of the landscape of such algorithmic private legal systems as it has developed to date. Building on prior work (Backer 2018, Backer 2021), the object is to theorize characteristics of emerging algorithmic systems and their distinctions among liberal democratic and Marxist Leninist political-economic models. The context centers on the examination of the landscape of such algorithmic private legal systems as it has developed to date by considering the extent to which a rating or algorithmic system has been emerging around recent national efforts to combat human trafficking through so-called Modern Slavery and Supply Chain Due Diligence legal regime and international norms.
We then turn to an exploration of the ramifications for the baseline premises against which the legitimacy of traditional legal models are measured. Among the more relevant are those tied to issues of privacy, of the integrity of data, and on transparency. More fundamental issues touch on the role of law in governance systems that are data driven and compliance oriented. One of the more interesting contradictions of the rise of algorithmic law is that societal stakeholders most committed to a critique of the use of the model on the basis of its threat to conventional protection of law (privacy, human rights, rule of law, democratic engagement and the like) are also likely to use algorithmic law to seek to harden human rights enhancing compliance systems based on data driven analytics with the characteristics.
We seek to advance the discourse of algorithmic law between two distinct lines of scholarship that have only recently emerged. The first is a line of scholarship that focuses its inquiry of algorithmic law and data-driven governance--by positioning the conversation around pragmatic issues, including the potential social harms and/or gains that could be had from algorithmic law ad more generally data driven governance (Campbell-Verduyn, Goguen and Porter 2016; Smith 2020; Katzenbach and Ulbricht 2019; Robinson 2017). In that vein some important scholarship has focused on issues of definition (Algorithmic Law) and quite influentially, others developed proposals on how to tame, contain and regulate their manifestation. (Pasquale 2017). Another line of scholarly development considered the ramifications of the rise of platforms used to support the structures and operations of algorithmic law (Barns 2019). Of particular utility have been scholarship of domestic governance policies supported using algorithmic law, and especially of the use of predictive analytics within multiple fields including governance, healthcare, economics, etc. (Van Claster, Wynants, Timmerman, Steverberg, Collins 2019; Lena, Delen 2020; Curran, Smart 2019). This discussion foregrounds ethics and the threat of algorithmic governance to established political values (for example privacy, autonomy, equal treatment; Pasquale 2015).
And yet, despite the threat especially to traditional liberal democratic values, algorithmic governance has great potential to overcome the obstacles to an effective development of a viable regulatory system for human rights protections in the context of economic activity. Ratings based regulatory structures swerve as a gateway for developing predictive analytics that has regulatory potential in ways that may serve liberal democratic values to the same extent as it appears now to serve Marxist Leninist values (Curran, Smart 2019). In this sense these mechanisms advances a discussion on algorithmic law’s role in international human rights law. (McGregor, Murray, and Ng 2019).
W. Richard and Mary Eshelman Faculty Scholar & Professor of Law and International Affairs, Pennsylvania State University
239 Lewis Katz Building, University Park, PA 16802 (1.814.863.3640 (direct), lcb11@psu.edu).
[2] School of International Affairs, Pennsylvania State University (MIA 2021).
[3] “Corporations, meanwhile, are increasingly adopting their own codes of conduct covering matters as diverse as environmental sustainability, labor rights, human rights, and standards of respect, honest and fair dealing with customers. These company-level norms are often enforced through processes that mirror the formal compliance function. Entities are being called on to encourage ethical and compliant behavior by third parties through systems such as programs of supply chain management,“ conflict minerals” disclosures, suspicious activities reports and similar activities (ALI 2019, Forward p. xv).
No comments:
Post a Comment