1 of 10

Responsible AI�- Welcome Remark -

2022.12.14

전길남

Kilnam Chon

KAIST

​

​

Korea AI Summit 2022: AI for Good 2022.11.18rev11.27

2 of 10

Responsible Science/Technology

Examples:

​

Responsible Nuclear Science/Technology

Nuclear Plant,…

​

Responsible Automobile Technology

Number of Fatalities per Year ~ 1.3 million

2

3 of 10

Cases in Korea

  • Ethics Class at KAIST CSD (and many other schools)

​

  • AI Governance Project at Seoul National University Law School (SAPI)

​

​

3

4 of 10

History in Korea

  • Internet

Following USA (too busy on everything else)

“Responsible Net”(?)

​

  • AI

Similar to Internet?

4

5 of 10

Ecosystem

  • Toward healthy ecosystem

With multistakeholders

5

6 of 10

References

S. Russell, Human Compatible, 2019.

S. Hawkins, Artificial intelligence could end mankind, BBC, 2014.

K. Chon, Governance of Digital Space, Asia Internet History, 2021

Kai-Fu lee, AI Superpowers, 2018.

6

7 of 10

Appendix: Open Issues

  • Internet AI

SNS (misinformation, abuse,….)

Security

  • Weapon

Game theory

7

8 of 10

Appendix Trustworthy AI Principles

  • Privacy Protection
  • Robustness/Stability
  • Interoperability
  • Fairness

​

​

​

​

[source: ChinAI News, 2022]

8

9 of 10

Appendix: Responsible AI Session

  • Opening
  • Welcome Remark
  • 7 Talks
  • Panel

9

10 of 10

10

Abstract: 

An overview on “Responsible AI” is given in this welcome remark to kick off the track.

The responsible AI is not easy to implement like many other science and technology issues. I give two examples: responsible nuclear science and technology including nuclear power systems, and  responsible automobile systems. Then, I show Korean cases; the ethics classes in computer science and AI governance conferences. I compare the histories of AI development and the Internet development in Korea next. Finally, I would like to mention on the AI ecosystem.  An healthy AI ecosystem may be the ultimate goal of the responsible AI.

�