When the United States moved in last month on Islamic State leader Abū Bakr al-Baghdadi al-Qurayshi, resulting in al-Baghdadi’s suicide by explosive vest, the U.S. conducted the military operation under the highest secrecy. The fact that the public learned so much detail about the mission after the fact was unusual.
The president, courts and Congress make national security decisions every day that the public may never hear about. There is little or no reporting, by design, because transparency could jeopardize U.S. national security and American lives.
So what level of accountability exists for classified decisions made in the interest of national security? Professor Ashley Deeks of the University of Virginia School of Law explores the issue in her new paper, “Secret Reason-Giving,” which is forthcoming in the Yale Law Journal.
Deeks, a national security law expert who formerly advised the State Department’s Office of the Legal Adviser and served as a U.S. embassy adviser in Baghdad, said the article developed from a series of faculty incubator lunches in which colleagues offered feedback about what she was working on.
At the first lunch, she told the faculty who joined her that she was interested in writing about the ways in which the use of artificial intelligence in national security decision-making could render today’s secret decision-making even less transparent. Military applications for AI might include facial-recognition software used to help identify a possible terrorist in disguise, or the monitoring of phone metadata that could predict a pending attack.
Deeks was fascinated by the implications of a machine helping to make a potentially fateful military decision, and how humans might explain or justify the decisions they make using those tools.
“A lot of national security decision-making is done behind closed doors, is done in a classified matter, and when you introduce artificial intelligence to military and intelligence processes, you are doubling down on the level of secrecy there, because it’s often not clear why machine-learning algorithms produce the outcomes they do,” she said. “So not only do you have a classified process, but you also have opaque tools helping you work through decisions in that classified process.”
This led her faculty peers to ask questions, she said: “How much do we really know about how the executive decides these national security questions in the first place? How often do executive officials really have to give reasons for those classified decisions?”
Deeks realized that she needed to get to the bottom of those questions before examining how AI might complicate the existing decision-making process.
During the second lunch, she discussed the different settings in which government officials provide their private rationales for national security-driven actions. Her preliminary research highlighted that all three branches of government sometimes engage in reason-giving in secret.
Among the better-known examples of secret reason-giving are those mandated by statute. The U.S. Foreign Intelligence Surveillance Act, for example, requires the executive branch to give the FISA court reasons why a foreign electronic surveillance warrant is merited. Likewise, a judge on the FISA court must give a written reason when denying a request, to facilitate the appeals process.
In other instances, such as a classified report prepared by the Senate Select Committee on Intelligence or an executive official making the case to Congress for conducting air strikes against an enemy, reasons would naturally follow in support of arguments and conclusions.
Checking executive branch decision-making is perhaps the most important function of secret reason-giving, Deeks argues in her paper. It “strikes a balance between two unappealing alternatives: allowing the Executive to decline to share its decision-making with any other branch and act unilaterally, or requiring the Executive to publicly share all of its decisions and justifications.”
Deeks maps in the paper how executive branch reason-giving can work internally among superiors, subordinates and peers inside the executive branch, and externally to the other two branches. She notes that secret reason-giving can even extend beyond the U.S. government to allies, as when the United States seeks to persuade those allies to join it in classified operations.
Although she said she believes the current presidency may be eschewing some of the traditional reason-giving not required by statute, the paper isn’t focused on a snapshot of one administration, but rather attempts to view reason-giving as part of internal executive norms that stretch across administrations.
She ultimately finds that “secret reason-giving improves the overall quality and effectiveness of government decision-making and operations, constrains the decision-maker, and strengthens the decision-maker’s legitimacy.”
At UVA, Deeks is E. James Kelly, Jr.–Class of 1965 Research Professor of Law; a senior fellow at the Center for National Security Law; and a faculty senior fellow at the Miller Center. She has written articles on the use of force, executive power, secret treaties, the intersection of national security and international law, and the laws of armed conflict. She is a member of the State Department’s Advisory Committee on International Law and serves as a senior contributor to the Lawfare blog. Deeks also serves on the boards of editors of the American Journal of International Law and the Journal of National Security Law and Policy. She is the supervising editor for AJIL Unbound and a senior fellow at the Lieber Institute for Law and Land Warfare.
Prior to academia, she served as the assistant legal adviser for political-military affairs in the U.S. State Department’s Office of the Legal Adviser, where she worked on issues related to the law of armed conflict, the use of force, conventional weapons, and the legal framework for the conflict with al-Qaida. She also provided advice on intelligence issues. In 2005, she served as embassy legal adviser at the U.S. Embassy in Baghdad during the nation’s constitutional negotiations.
Deeks received her J.D. with honors from the University of Chicago Law School.
“Behind the Scholarship” is an occasional series that details the backstories of the faculty’s work.
Read More
- “Secret Reason-Giving,” Yale L.J. (forthcoming 2020).
- “The Judicial Demand for Explainable Artificial Intelligence,” Colum. L. Rev. (forthcoming 2019).
- “Machine Learning, Artificial Intelligence, and the Use of Force by States” (with Noam Lubell and Daragh Murray), J. Nat’l Security L. & Pol’y (forthcoming 2019).
- “Predicting Enemies,” 104 Va. L. Rev. 1529 (2018).
- “The Substance of Secret Agreements and the Role of Government Lawyers,” 111 AJIL Unbound 474 (2018).
- “An International Legal Framework for Surveillance,” 55 Va. J. Int'l L. 291 (2015).
- “Unwilling or Unable: Toward a Normative Framework for Extra-Territorial Self-Defense," 52 Va. J. Int’l L. 483 (2012).
Founded in 1819, the University of Virginia School of Law is the second-oldest continuously operating law school in the nation. Consistently ranked among the top law schools, Virginia is a world-renowned training ground for distinguished lawyers and public servants, instilling in them a commitment to leadership, integrity and community service.