AI: Intelligent machines, smart policies

20 Dec 2017 16:45h - 18:15h

Event report

[Read more session reports and live updates from the 12th Internet Governance Forum]

The session began with moderator Mr Wonki Min, Deputy Minister, Ministry of Science, ICT and Future Planning, Republic of Korea, welcoming the attendees and explaining that the panellists would focus their presentations exclusively on issues related to the policy agenda on Artificial Intelligence (AI). In addition, he noted that the organisation of the workshop was possible thanks to the support of the Ministry of the Interior and Communications of Japan, which had initially expressed interest in an AI event held in Paris in 2016.

The first panellist was Mr Masahiko Tominaga, Vice Minister, Internal Affairs and Communications, Japan, who indicated that Japan is currently leading discussions on AI and is developing “AI networking” which means connecting AI to different devices via the Internet. However, he pointed out that many challenges have been encountered, not only technical but also questions about the transparency, and risks around use and control of these technologies.

Tominaga also mentioned that the discussion on the development of AI has been taking place in different forums. In the case of Japan, he indicated that since 2016 there has been an Advisory Committee within the Ministry of the Interior and Communications to evaluate the social, economic and political impact of AI. This Commission has developed a proposal related to the impact of the technology and in addition to assessing the opportunities, it seeks to mitigate risks, especially in terms of job creation and the changes that must take place in education and capacity building. He concluded by indicating that the Commission has defined some principles about AI: Declaration, transparency, controllability, safety, security, privacy, ethics and accountability.

The next panellist was Ms Anne Carblanc, Head of Digital Economy Policy Division, OECD, who cited some current initiatives around AI. She indicated that organisations such as the OECD are currently looking for where the initiatives on AI are, how they are being developed, and how their socioeconomic impact is measured in order to develop best practices in this field.

Carblanc then referred to an event in which experts and government entities were invited to participate, not only in the area of ​​public policies related to the digital economy, but also other areas such as labour, and data protection. Findings of this event included the verification that AI was perhaps the technology which had generated more profound and immediate impacts than any other technology in such a short time. Salaries, unemployment, and privacy, were also of interest to the participants.

The next panellist was Ms Joanna Bryson, Reader at the University of Bath, and Affiliate, Center for Information Technology Policy at Princeton University, who began by defining what is meant by “intelligence”. Intelligence, she said, is doing the right thing at the right time. This simple concept is not as simple as it seems, insofar as it involves making many complex decisions that have to do not only with prior knowledge but also with the context of those decisions.

Bryson also indicated that it should be understood that this type of intelligence is limited by time and space. Time, since it is possible that a computer can eventually solve a problem, but in a time so big that the answer may have lost its sense. In the case of space, she used the example that there are more movements in chess than atoms in the universe, but if a computer wanted to measure them, where would it place all that information?

Finally, Bryson indicated that she disagrees with some ideas related to AI, especially those that argue that many humans will be replaced by this technology. She said that one of the things that we know in biology is the more variation you have, the more robust you are. So where does innovation come from? It comes from having a lot of people that have very different education, very different backgrounds. This explains why a technology made by humans should not try to replace them.

The next panellist was Ms Carolyn Nguyen, Director of Technology Policy, Microsoft, who spoke about what the development of AI promises for the world. Some of these promises are: That economic growth will double by 2035, that the Sustainable Development Goals (SDGs) will be better addressed, and that governance processes will be strengthened in areas such as health.

Nguyen added that Microsoft believes that the development of AI should not be governed by ethics, which depends a lot on the society where this is defined, but should be based on principles such as trust, justice, security, privacy, etc. She also indicated that there is currently a group, Partnership on Artificial Intelligence to Benefit People and Society, which seeks to raise these issues and brings together people from all over the world, in an attempt to reinforce collaboration as the axis of development of AI.

Ms Karen McCabe, Senior Director, Technology Policy and International Affairs, IEEE Standards Association, announced that within the IEEE there have been more than 13 working groups on AI since 2016 that bring together more than 300 experts from different countries. McCabe said that there is great interest in the subject, and some groups include discussions on ethics and policy making, always oriented to the development of human well-being.

McCabe also said she believes it is convenient to regulate this technology in order to avoid the risks and potential effects on people. The goals of an effective AI policy should centre around the protection and promotion of safety, privacy, and intellectual rights, as well as the impact of these systems on society, she said.

The last panellist was Mr Jean-Marc Rickli, Global Risk and Resilience Cluster Leader, Leadership, Crisis and Conflict Management Programme, who presented an initiative on AI for society, arguing that it was very important that everyone get involved in the development of this technology. The launch of the programme was in September and work is ongoing to define basic work principles.

Rickli explained that the objective was to be able to discuss the substantive problems that currently exist and which are related to AI. For that, some initial objectives have been created: Reinventing the machine, security, governance of AI, workforce for the age of AI, AI for the public good, and images of AI. He indicated that all this information can be currently consulted on the website of the initiative.

After the presentations, a round of questions was opened in which the participants could make comments and clarifications. The majority dealt with the little information that exists about AI in countries of the “Global South” and about the level of development of this technology in order to know what to expect in the immediate future. At the end of the round, Mr Wonki Min thanked everyone for their attendance and concluded the session.

By Carlos Guerrero