International Lawyering in the High-Tech Era
Ashley Deeks - School of Law, University of Virginia
-- It’s 2025. Imagine that the United States and Russia decide to negotiate an extradition treaty. Newspapers in both countries have been reporting on the impending negotiation, and the public has been discussing it on social media. Russia already has fifty extradition treaties with other states. U.S. international lawyers and data scientists decide to train machine learning algorithms on the texts of these agreements to identify what the “average” Russian extradition treaty looks like, extract the most common (and most exceptional) terms in the existing treaties, and enter the negotiation well-positioned to anticipate Russian requests. The U.S. negotiators also mine information from Russian social media to help them understand public sentiment and see that the draft treaty provision about political offenses is likely to be contentious. The U.S. side drafts the proposed provision to take account of that controversy. Later, during the negotiations, the U.S. team trains concealed video cameras on the Russian negotiators. During a break, the U.S. team runs emotion recognition algorithms over the videos to identify the parts of the negotiation during which the Russian team acted deceptively and adjust their strategy accordingly.
International lawyers who work for governments do not use these technologies today. Indeed, international law generally has been a stranger to a new wave of technological tools – including computational text analysis, machine learning, and predictive algorithms – that use large quantities of data to make sense of the world. But actors in a wide range of other fields use data-driven algorithmic tools to improve medical diagnoses; enable greater autonomy in daily activities such as driving; enhance the accuracy of facial recognition software; facilitate inter-personal negotiations; and improve decision-making. Further, these tools have proven valuable in domestic legal practice, negotiations, and other fields closely related to international law.
In a forthcoming article, "High-Tech International Law," I examine how international lawyers might take advantage of these high-tech tools. Faced with large volumes of information (all of a state’s submissions ever made to the United Nations, for instance, or thousands of arbitral judgments by particular arbitrators), a foreign ministry lawyer with limited time and resources could deploy algorithmic tools to help her extract important insights from that data, improving her negotiation posture or allowing her to provide more accurate predictions to a client about how an international arbitral panel will rule. The article is not intended to suggest that humans soon will, can, or should turn over these types of international activities wholesale to machines. Rather, it suggests that there can be significant benefits to allowing machine learning tools to support human decision-making in these areas.
High-tech international law has the potential to significantly alter the distribution of power among states. Any evaluation of international law-related technologies should therefore also consider who the winners and losers will be as this technology gains a foothold in international law processes such as treaty negotiation and dispute resolution. Here, the story appears mixed: some of these tools will empower actors in less powerful states, improving their access to information and their ability to use that information instrumentally. Other tools will empower the already-powerful, further challenging the fiction of sovereign equality. There are steps that less powerful states and outside actors can take, however, to counter-balance power shifts that could hinder, rather than advance, the ambitions of international law.