ICT4Peace and ZHET Panel Discussion at RightsCon Toronto 2018 on Artificial Intelligence: Lethal Autonomous Weapons Systems and Peace Time Threats with:
- Todd Davies, Associate Director and Lecturer, Symbolic Systems Program, Stanford University, USA
- Ron Deibert, Director, Citizen Lab, Munk School of Global Affairs, University of Toronto, Canada
- Kyle Dent, Author and Researcher, Palo Alto Research Center, PARC, USA
- Maarten Van Horenbeeck, Board Member and former Chairman, Forum of Incident Response and Security Teams (FIRST), USA
Moderated by:
David Kirkpatrick, Founder, Techonomy Media, and author, The Facebook Effect, USA
Dr Seeta Peña Gangadharan, Assistant Professor, Department of Media and Communications, London School of Economics, had to forgoe her participation as a panellist on very short notice for personal reasons.
The video recording of this panel discussion can be found here: https://www.youtube.com/watch?v=IWhUBSSEYk8
What is the risk that the tech that governs and determines modern society could itself become ungovernable? Will this be the means through which we ultimately lose control of society and democratic values? What impact does new tech and AI have both on warfare and our day to day lives? These were some of the questions raised by David Kirkpatrick during the ICT4Peace and ZHET panel on Artificial Intelligence: Lethal Autonomous Weapons Systems and Peace Time Threats at RightsCon in Toronto on 18 May 2018 .
There was broad agreement that we need to urgently educate engineers and techies on ethics and ethicists on tech as stressed by Kyle Dent. We also need to change the incentives for engineers who will, without other guidance, always aim for a tech solution. Algorithms that are currently, often accidentally, programmed with bias need to be designed with values supporting human rights and protecting individuals. We need to safeguard the quality of the information and data being used in developing AI and the underlying algorithms.
The protection of researchers was highlighted by Ron Deibert who argued strongly for measures to protect academic freedom, citing the legal issues CitizenLab has faced with Netsweeper and Sandvine. We live in an economy that is a personal data surveillance economy in which companies are harvesting and mining as much data as they can about us. How can we ensure this does not take us down the slippery slope of invading personal freedoms and jeopardizing the democratic values and principles of human rights that have been won with great difficulty through the course of history? Technological change does not happen in a vacuum and we need to take conscious decisions about the type of world we would like to live in.
The challenge of AI in Lethal Autonomous Weapons Systems was also addressed by Todd Davies, who highlighted the Campaign to Stop Killer Robots, and the 2015 Open Letter from AI researchers including Stephen Hawking and Elon Musk. LAWS are dangerous for many reasons including our inability to defend against them, the fact that a small number of people could deploy large numbers of LAWS, they are relative cheap and based on software technology that is vulnerable to hacking and accidental use. They are also not accountable before the law.
Finally, as Maarten Van Horenbeeck noted AI is at the core of many security technologies. Identifying what is causing an error or where an attack might be coming from will become even more difficult over time. The intersection between cybersecurity and AI needs to be assessed and is of increasing importance. We need to push for explainable AI and the auditing of AI and move away from the black box problem that comes with deep learning and opaque technology. As we build more and more complex algorithms we may get to a place where turning them off and doing things manually might not work anymore. This is not in anyone’s interest.
The ICT4Peace paper on AI, Lethal Autonomous Weapons and Peace Time Threats by Regina Surber, Senior Advisor, ICT4Peace, served as a background paper for this session and can be found here: 2018_RSurber_AI-AT-LAWS-Peace-Time-Threats_final. The recording of her lecture on the same topic can be found here: https://www.youtube.com/watch?v=dDJ0uUqz_go
The Panel was designed and prepared by Barbara Weekes, Senior Advisor, ICT4Peace Foundation.