IGF 2017 WS #91 Policy Challenges for AI Development

Short Title: 

Policy Challenges for AI Development

Proposer's Name: Mr. Rui ZHONG
Proposer's Organization: Internet Society of China
Co-Proposer's Name: Ms. Ping Wu
Co-Proposer's Organization: Internet Society of China
Ms. Jing MA, Civil Society, China Association for Science and Technology

Additional Speakers: 

Ms. Akari Noguchi, worked for public policy & corporate governance,Yahoo Japan; Mr. Cláudio S. de Lucena Neto, Visiting Researcher of Georgetown University, Professor of Law, Center for Legal Studies, Paraíba State University of Brazil; Ms. Jia HE, Person in Charge of Think-Tank and Corporate-Social-Responsibility Affairs, Bytedance (China);


1. Xiaoyan Ran, Internet Society of China (3 mins; Opening remark) 2. Akari Noguchi, Yahoo Japan (8 mins; Topic: Policy for AI's development) 3. Cláudio S. de Lucena Neto, Visiting Researcher of Georgetown University (8 mins; Topic: challenges of policy development) 4. Satish Babu, Chair of ICANN APRALO (8 mins; Topic: Ethical aspects of AI Policy) 5. Yi Ma, Professor of Electrical Engineering and Computer Sciences, UC Berkeley. (8 mins; Topic: AI policy challenges) 6. Discussion, on-site and remote interactive Q&A (45 mins) 7. Xiaoyan RAN, Internet Society of China (2 mins; Closing remark for the workshop)


- Session Title:   Policy Challenges for AI Development             
- Date:  Dec 19, 2017                
- Time:  09:00—10:30               
- Session Organizer:  Internet Society of China             
- Chair/Moderator: Xiaoyan RAN             
- Rapporteur/Notetaker: Yexin LIU           

- List of Speakers and their institutional affiliations:     

1.Mr. Satish Babu, Chair of ICANN APRALO, Founding Director of the International Centre for Free and Open Source Software

2.Ms. Akari Noguchi, worked for public policy & corporate governance in Yahoo Japan

3.Mr. Cláudio S. de Lucena Neto, Visiting Researcher of Georgetown University, Professor of Law, Center for Legal Studies, Paraíba State University of Brazil

4.Dr. Yi Ma, Professor of Electrical Engineering and Computer Sciences, UC Berkeley; Fellow of IEEE and ACM.

Key Issues raised (1 sentence per issue):            

  • How fast AI industry develops;
  • Ethical issues should be considered ahead when we make AI policies;
  • AI can help enterprises to develop more and take Yahoo Japan for example;
  • The roles which non-profit institutions can play in terms of AI development;
  • What are the suggestions for AI policy-making.

- If there were presentations during the session, please provide a 1-paragraph summary for each presentation:             

  • Mr. Satish Babu gave words themed as Ethical aspects of AI Policy, such as whether robots should have rights like human beings do, including privacy; or how we should handle the unemployment issue when AI is applied. With no regard to ethical aspects of these, we may result in having AI and AS(artificial stupidity) at the same time.
  • Ms. Akari Noguchi introduced how AI helped Yahoo Japan in shopping recommendation, behavior targeting by past behavior analysis and modeling in advertising, spam mail filtering, fostering security system to monitor and suspend the fake login, etc. Using AI to find new patterns of fraud by current Machine learning and deep learning would be their plan for 2018.
  • Mr. Cláudio S. de Lucena Neto shared his thoughts on who (institutions that were currently involved in policy debate), what was in their agenda and how they were doing it, and what was yet to be done in terms of policy development. He also wrapped up with the challenges he found the most difficult to overcome and emphasized on the import role IEEE was taking.
  • Dr. Yi Ma believed that AI policy challenges, divided into three aspects, were—policies should accommodate differences of markets and nations; policies should resolve concerns with security, safety and privacy; policies should encourage companies to do good for the society. Great progress had been made both in AI industry and technology, and many changes had taken place promoted by private companies like certain social media had applied Bytedance (a Chinese company powered by AI) program for finding missing people with good results. So policy-making needs to catch up.

- Please describe the Discussions that took place during the workshop session (3 paragraphs): 

  • The discussions took place right after the presentations and about 5 participants shared their viewpoints or raise some questions for the speakers.
  • The questions were answered ardently by not only the speakers but also the audience. While answering the former questions, new issues were brought up. The discussions came to an end pitifully because time was running out so many participants waited to share a few words with the speakers after the workshop.
  •  Main issues discussed were how different groups like programmers and businessmen could  solve their differences quickly in order to march towards the same goal of developing AI; what were the similarities between AI and other computing science like software programming; how to deal the unemployment issue, etc.

- Please describe any Participant suggestions regarding the way forward/ potential next steps /key takeaways (3 paragraphs):    

  • To master the challenges for AI development, we should pay attention to determining its responsibility, accountability and maturity as much as we can.
  • There is no quick way to stimulate different stakeholders to devote to AI cause, and the most important thing remained ‘dialogue’, which means communication among any group involved. No one could manage achieving AI development by oneself.
  • NGO institutions, international or domestic, possessed great potentials for uniting different stakeholders to develop AI, many of which had already done many great jobs.

-Gender Reporting

1.- Estimate the overall number of the participants present at the session:  150
2.- Estimate the overall number of women present at the session: 75
3.- To what extent did the session discuss gender equality and/or women’s empowerment? Very little
4. -If the session addressed issues related to gender equality and/or women’s empowerment, please provide a brief summary of the discussion:None


Session Format: Panel - 90 Min

Country: China
Stakeholder Group: Civil Society

Country: China
Stakeholder Group: Civil Society

Speaker: Alison Gillwald
Speaker: Satish Babu
Speaker: Kenta Mochizuki
Speaker: Lei Ning
Speaker: Shenshen Cao

Content of the Session:
The application of AI will be quite extensive, for instance in the field of digital advertising, agriculture, sales, even music, art, psychology and charity. However, the development of AI will encounter many challenges, such as policy issues on data protection, ethical considerations in AI's design core value and neutrality. Therefore, some sorts of approaches and issues are necessary to be further discussed for the development and governance of AI.

This workshop will invite multi-stakeholder representatives to discuss what policy challenges AI will be encountered and how to create an enabling policy environment that conducive to the sustainable development from different dimensions and aspects.

Intended Agenda and issues to be discussed:

1.Xinmin Gao, Internet Society of China (3 mins; Opening remark)
2.Kenta Mochizuki, Yahoo Japan (8 mins; Talk about user's AI data privacy and protection)
3.Alison Gillwald, University of Cape Town (8 mins; Talk about ICT technology and policy advice for government)
4.Satish Babu, IEEE (8 mins; Talk about AI's challenges in Internet governance from perspective of technology)
5.Lei Ning, Baidu (8 mins; Talk about AI's challenges in Internet governance from perspective of product and international cooperation)
6.Shenshen Cao, Tencent (8 mins; Talk about AI's challenges in Internet governance from perspective of public policy and industry's sustainable development)
7.Discussion, on-site and remote interactive Q&A (45 mins; we will try to invite other potential guests to join discussion, such as Google, Microsoft)
8.Xinmin Gao, Internet Society of China (2 mins; close the workshop)

Relevance of the Session:
Artificial Intelligence(AI)is considered as one of important rising ICT industries and is shaping the digital future of human being. Technically, AI basically involves with Machine Learning, which means it will potentially collect, analyze and use large amounts of data, particularly some of which combined with other information could be identified to one’s personal information. In accordance with the legal provisions of personal information protection, these acts should be clearly, fully and completely authorized by the users, and also the purposes, means, content, retention duration and use scope etc. should be clearly informed to the users. Therefore, what kinds of specific policies and measures necessary in the process of collecting and using data should be discussed so as to guide and instruct the AI developer obeying the security principle, taking appropriate management measures and technical means based on the possibility and severity of personal information damage, as well as effectively protect personal information from unauthorized retrieval, loss, disclosure, damage and tampering.

Furthermore, as the machine is designed by the developer, the program naturally endowed with the designer's value orientation, which means the developer need to uphold a wide range of inclusiveness in the AI training and design process, fully considering the interest of women, children, the disabled, minorities and other vulnerable groups, as well as set special rules of judgment in extreme morality and law circumstance. In that the AI system is not as "technically neutral" as it looks like, the specific group may turned out to be the victims of systematic "prejudice" and "discrimination" unwittingly.

Generally, research and development of AI technologies or products are not restricted by administrative license and access. However, once these technologies and products applied to specific industry, there may arouse industry license issues. Will the industry regulation extend to AI field in the future? In finance, medicine, smart home, automatic driving and other specific fields, will regulation be necessary to intervene in the AI development? For example, if a driverless car with passengers encountered an accident, who should bear the main responsibility and how to monitor such situation?

Thus, how to govern the AI through practical policy approaches is very vital consideration for many areas, such as cyber security, personal information protection, network neutrality, ICT for sustainable development.

Tag 1: Digital Future
Tag 2: Internet of Things
Tag 3: Internet & ICTs for the Sustainable Development Goals

As the diverse background and knowledge difference of the speakers, we will invite them to share insights from different issue angles, including policy making advice, technical practice, product development, industry research etc.

In the current stage, we have five speakers from different stakehloder groups (private sector, technical community and civil society), and two female speakers. Most of them come from developing countries and some of them are first time participating in IGF.

Onsite Moderator: Xinmin GAO, VP, Internet Society of China
Online Moderator: Rui ZHONG, Director, Internet Society of China
Rapporteur: Rui ZHONG, Director, Internet Society of China

Online Participation:
The workshop is open to every one, both onsite and remote participation. We will have a remote moderator to interact with the remote participants and will foward their questions and insights to the workshop speakers and audience. We will post the message of workshop info online in advance to let more people knowing what will be discussed and welcome joining from the global.

Discussion facilitation:
The workshop will be set as panel format, firstly allowing each panelist to share her/his main views and then have mutual discussion about other speakers' viewpoints. The on-site moderator also raise some questions prepared. To listen to more voice, we reserve more time slot for audience and remote participants to ask questions and make comments.

Conducted a Workshop in IGF before?: Yes
Link to Report: http://www.intgovforum.org/cms/wks2014/index.php/proposal/view_public/41