In this PHAP Expert Legal Briefing, Naz Modirzadeh and Dustin Lewis, from the Harvard Law School Program on International Law in Armed Conflict, explored developments in technology, accountability, and international law pertaining to armed conflict. The background concern is that in war, as in so many areas, power and authority are increasingly expressed algorithmically. Advancements in artificial intelligence and robotics may implicate—and possibly transform—numerous aspects of armed conflict. For instance, increasingly sophisticated forms of technical autonomy may affect the conduct of hostilities (including the development and use of “autonomous weapons”). But they also might relate to other elements pertaining to war, such as guarding and transporting detainees, providing medical care, and delivering humanitarian assistance.
 
The presenters summarized a recent report from the Harvard Law School Program on International Law and Armed Conflict. That report introduces a new concept—war algorithms—that aims to elevate algorithmically-derived “choices” and “decisions” to a central concern regarding technical autonomy in war. The report defines a “war algorithm” as any algorithm that is expressed in computer code, that is effectuated through a constructed system, and that is capable of operating in relation to armed conflict. Through the “war algorithms” lens, the presenters linked international law and related accountability architectures to relevant technologies.

Read the interview with Dustin A. Lewis on the emergence and relevance of this topic

Event recording (YouTube - faster loading time)

Event recording (Adobe Connect - better quality)

Event recording (audio podcast)

Target audience

The Expert Briefings specifically target legal experts and humanitarian practitioners needing an advanced legal knowledge.

Recommended resources

• Dustin A. Lewis, Gabriella Blum, and Naz K. Modirzadeh, “War-Algorithm Accountability,” Research Briefing, Harvard Law School Program on International Law and Armed Conflict, August 2016.
    - Executive Summary
    - Bibliography
    - Summary of state positions on “lethal autonomous weapons systems”
• Gabriella Blum, Dustin Lewis, and Naz Modirzadeh, “Accountability for Algorithmic Autonomy in War,” Lawfare Blog, September 12, 2016.
• Rebecca Crootof, The Varied Law of Autonomous Weapon Systems, in Autonomous Systems: Issues for Defence Policymakers (Andrew P. Williams & Paul D. Scharre eds., 2015).

Further resources

During the event the following resources, websites, and videos were mentioned by speakers and participants:

• ‘Introducing SpotMini’ video
• ‘Super aEgis II’ video
• ‘K-MAX Unmanned Helicopter’ video
• ‘A Swarm of Nano Quadrotors’ video
• ‘Stop Killer Robots’ video
• Stop Killer Robots, campaign's website
• Blog on Killer Robots by Harvard Human Rights Program
• John Reed, “The campaign to ban killer robots is here,” in Foreign Policy, 19 November 2012.
• Matthew Rosenberg and John Markoffoct, “The Pentagon’s ‘Terminator Conundrum’”, in The New York Times, 25 October 2016.
• Michael N. Schmitt and Jeffrey S. Thurnher, "'Out of the Loop: Autonomous Weapon Systems and the Law of Armed Conflict," in Harvard National Security Journal, vol. 4, 2013.
• Defence Science Board, "Summer Study on Autonomy," US Department of Defense, June 2016.
• AlphaGo Games, to learn more about these games.
• Christof Koch, "How the Computer Beat the Go Master," in Scientific American, 19 March 2016.
• David Silver et al., "Mastering the game of Go with deep neural networks and tree search," in Nature, vol. 529, issue 7587, 28 January 2016. (Full version requires paid access)
• Yann LeCun, Yoshua Bengio, and Geoffrey Hinton, "Deep learning," in Nature, vol. 521, issue 7553, 25 February 2015. (Full version requires paid access)

In addition, the following courses were recommended in order to learn more about algorithms, artificial intelligence and machine learning:

• MIT course with video lectures available on Introduction to Algorithms (SMA 5503 at MIT, Fall 2005).
• MIT course with lecture notes available on Design and Analysis of Algorithms (6.046J / 18.410J at MIT, 2012).
• MIT course with video lectures available on Artificial Intelligence (6.034 at MIT, Fall 2010).
• A Visual Introduction to Machine Learning from R2D3.
• Standford University course on Machine Learning at Coursera.
• University of Toronto course on Neural Networks for Machine Learning at Coursera.

Speakers

Event host

Angharad Laing Executive Director, International Association of Professionals in Humanitarian Assistance and Protection (PHAP)