This blog post grows out of the author’s recently published paper entitled “Will the ‘Legal Singularity’ Hollow Out Law’s Normative Core?”
Predictive analytic technologies have made dizzying advances in recent years. Powered by increasingly powerful and effective machine learning that trains itself on datasets that were unthinkably large even a decade ago, these technologies are already changing the workaday toolkit of those who interface with the legal system. Machine learning algorithms are already outperforming human lawyers at predicting the resolution of cases before the Supreme Court and the European Court of Human Rights. Litigators are using practical applications such as Lex Machina and Ravel to generate probabilistic point estimates of case outcomes. Last year, Estonia rolled out a system of AI judges that—or is it “who”?—are issuing binding legal judgments in small claims cases.
Legal futurists and other techno-enthusiasts are licking their chops. Two economists have already proposed to fully algorithmizing the judicial system, populating it with “digital judges” running on Blockchain software. Why? For the sake of reducing “high judicial costs”, of course.
Never mind that, in the U.S. at least, the budget of the federal judiciary last year was only 0.033% of national gross domestic product. An uncompromising neoliberal logic of efficiency and optimization, laser-focused on cost reduction, motivates most of these projects. Indeed, some legal futurists and techno-enthusiasts herald these early-stage interventions as merely a preface to the eventual arrival of a legal singularity, defined as the moment when predictive, mass-data analytic technologies evolve to create a perfectly predictable, algorithmically-expressed legal system bereft of all legal uncertainty.
The exact trajectory predictive legal analytics will take in the coming years is really anyone’s guess, but we can be fairly sure that the discourse of cost cutting and increased rationality via technology will be coming for the legal system, piece by piece. As such, it is worthwhile for left legal theory to take legal futurism, and even the idea of a legal singularity, seriously.
Right now, an ongoing robust debate among progressive legal scholars is highlighting the negative social externalities resulting from the deployment of predictive analytic technologies within the legal system. The harms to which this research program directs our attention are largely external to the legal system—effects on economic security and inequality, personal liberty, privacy, and so forth.
But less examined are predictive legal technology’s internal effects on the legal system—namely, the unintended jurisprudential consequences of a hurried embrace of legal futurism that threaten to destabilize the basic self-conception of liberal legal systems. In particular, an uncritical embrace of legal futurism would undermine the rule of law principle, a bedrock institution of any self-described liberal legal order and a key source for the legal system’s normative legitimacy. It will do so by dissolving the normative content of its two core constituents: predictability and universality, each of which has traditionally been conceived as a bulwark against arbitrary government power.
The predictability principle of the rule of law states that a liberal legal system must ensure that affected legal subjects can predict how the law will apply to them. Futurists claim to value predictability—indeed, the ability of AI to efficiently predict outcomes doubles as its core value proposition and its primary basis for legitimacy. But this form of predictability is an unacceptably weak version. Traceable to Hobbesian political theory, it prizes predictability because it enables subjects to plan their affairs and thereby fosters social stability. This is perhaps a fine principle for business owners and Holmesian “bad men,” but it is too thin a principle to ground a legal system that values the consent of the governed. If predictability is to be valuable to that end, it must be comprehensible, intelligible, and amenable, at least in theory, to contestation. As Locke put it, the authorities must “own willingly” their exercise of governmental authority, making it possible for the citizenry to demand reasoned explanations for incursions in their otherwise natural rights to liberty and autonomy.
The futurists simply ignore the institutional importance of predictability in this strong sense. They conceive of the legal system as a ready-formed thing that they would like to make susceptible to predictive analysis—as a data set ripe for empirical investigation, not as a reflexive social process, embedded in real social institutions, that relies for its normative legitimacy on its scrutability and susceptibility to review and contestation. The normative force that the rule of law imparts to the legal system originates in this strong form of predictability, and the legal futurists ignore it altogether in favor of a brute positivism that treats the law instead as an analytically cleaner, but normatively empty, system.
Legal futurism also threatens to destabilize the rule of law’s universality principle by reconceiving of legal subjects as aggregations of data points rather than as individual members of a polity. The issue here is that the basic idea of universal application of law might become unintelligible in the face of an epistemological shift that allows a newly algorithmized legal system to only see pulverized, atomized data points where it used to see integral, individual legal subjects. Consequently, legal futurism stands to undermine the universality principle’s premise that the differences among legal subjects are outweighed by what we—or, better still, “We the People” who are, as Blackstone put it, the “community in general”—have in common.
In failing to live up to these principles, the prospect of an algorithmically-generated, futurist legal system will be shot through with arbitrariness and untethered to conceptions of citizens’ welfare and consent. It will be an “algocracy,” to borrow John Danaher’s helpful term. Such an algocracy would present a choice. Since you can’t simultaneously prize the rule of law and liberal legalism and embrace techno-futurist visions of algorithms replacing human legal judgment and institutions, legal futurists must determine whether to jettison the fundamental liberal values of law or their hopes of purely algorithmic governance.
But this sounds like a problem for liberals. Why should progressives, let alone radicals, care about the dissolution of the rule of law, or wanting to reinvigorate legal liberalism? Isn’t the primordial principle of left legal theory that liberal legal systems operate as instruments/expressions/condensations of dominant class interests?
Even if we agree with that perspective as a general background condition, it’s possible to reappropriate liberal concepts for more radical purposes, recognizing that, for better or worse, liberalism is a category that modern political and legal thought simply cannot escape. Leftist political and legal theorists such as E.P. Thompson, Franz Neumann, and Jürgen Habermas have acknowledged the importance of the rule of law in spite of its ambivalent status as a tool of dominant class interests. And they have compellingly warned about the potential for crisis in societies in which the rule of law disappears.
Indeed, given the increasingly awkward dance between liberalism as a philosophical program and neoliberal projects in their existing political and institutional forms—of which legal futurism, with its underlying logics of optimization, entrepreneurialism, scarcity, and efficiency, is a signal example—it might very well be that any attempt to save legal liberalism from techno-enthusiastic excess must come from the legal left. This is because progressive legal theory can straightforwardly marshal the normative energy to criticize legal tech without the apparent contradictions with which liberals-cum-neoliberals must wrestle. Whereas the latter might perceive a tradeoff between the neoliberal imperative of cost minimization and a liberal commitment to nonarbitrary government, progressives clearly do not.
By unmasking the contradictions between legal futurism and legal liberalism, we can contribute to efforts to develop a new liberalism, along the lines of what the philosopher Charles Mills calls an emancipatory, radical liberalism. According to this approach, the problems with liberalism are not immanent to its conceptual logic (egalitarianism, universalism, freedom, autonomy, individual dignity), but instead are the result of material group interests advocating for hierarchical, exclusionary, myopic versions of liberalism. The shrunken form of predictability in futurist visions of the legal system is, I think, just such a perspective, justified as it is by the logics of efficiency, optimization, and scarcity with no regard for the normative fallout that will result.
Reinvigorating radical liberal theory in this way can help light the way for practice, in the form of a new radical politics of law and technology committed not to preserving the status quo liberal legal system against technological intrusions, but to breathing new normative life into increasingly technologized legal institutions. The most pressing practical task for such a politics must be to imagine and design the new legal institutions responsible for counteracting the threat of algorithmic arbitrariness in government. In performing that work, we should lean on transparency, precaution, and democratic participation that operates across class, race, and gender.
Who knows: as we pursue this task, we might even enjoy the company of some authentic liberals, nostalgic for a meaningful conception of nonarbitrary government and the consent of the governed, who will join us as co-belligerents against a common algocratic threat.
This radical politics of law and technology should also be seen as part of the broader confrontation between left social theory and the dominant socio-technical ecosystem. This “technosystem” is rapidly and unremittingly transforming our social world, without any meaningful input or consent from those whose lives it increasingly dominates. Technology critic Adam Greenfield has argued that this era of radical technologies demands a generation of radical technologists. Radical legal scholars and lawyers should expect that these radical technologists will need our help to advocate in legislative and (state and non-state) regulatory settings, and to navigate the questions of institutional design, finance, and governance that any radical politics of technology will inevitably generate. We should prepare ourselves for that call.