i Origins February 24 — 26, 2017 PROJECT An Origins Project Scientific Workshop ARIZONA STATE UNIVERSITY Challenges of Artificial Intelligence: Envisioning and Addressing Adverse Outcomes systems. Effective meaningful human control faces challenges with the interpretation and fusion of sensor signals and the understanding of humans of Al pattern recognition and inference. DISCUSSION What methods, international norms, agreements, communication protocols, regulatory activity, etc. might be harnessed to minimize challenges with destabilizations around time-criticality, automation, and gaming? How can meaningful human control be assured/inserted into key aspects of decision making? REFERENCES Report Cites Dangers of Autonomous Weapons, New York Times, Feb. 28, 2016. https://www.nytimes.com/2016/02/29/technology/report-cites-dangers-of-autonomous- weapons.html The Morality of Robotic War, New York Times, May 27, 2015 https://www.nytimes.com/2015/05/27/opinion/the-morality-of-robotic-war.html P. Scharre, Autonomous Weapons and Operational Risk, Center for a New American Security, February 2016. https://s3.amazonaws.com/files.cnas.org/documents/CNAS_Autonomous-weapons-operational- risk.pdf US Department of Defense Directive 3000.09, November 21, 2012 http://www.dtic.mil/whs/directives/corres/pdf/300009p. pdf P. Scharre and M.C. Horowitz, An Introduction to Autonomy in Weapon Systems. https://s3.amazonaws.com/files.cnas.org/documents/Ethical-Autonomy-Working- Paper 021015 v02.pdf 9 HOUSE_OVERSIGHT_014705