NAVER to Hold Workshop on Hyperscale AI Ethics at AI International Conference
NAVER to Hold Workshop on Hyperscale AI Ethics at AI International Conference
NAVER to Hold Workshop on Hyperscale AI Ethics
at AI International Conference
- To hold workshop entitled ‘Fairness, Accountability and Transparency in Hyperscale Language Models’ during ACM FAccT
- AI ethics experts from academia and industry, including Professor Ko Hak-soo of Seoul National University, Professor Cho Kyung-hyun of New York University and Margaret Mitchell, to participate in the event
2022-06-22
NAVER CLOVA participated in the ACM Conference on Fairness, Accountability and Transparency (ACM FAccT), a leading international conference on computer science integrity. At the event, NAVER CLOVA held a workshop and a tutorial session on the topic of hyperscale AI-related ethics with world class experts. NAVER plans to secure global leadership in the field of artificial intelligence (AI) ethics while leading the development and application of hyperscale AI technologies through NAVER’s hyperscale AI HyperCLOVA.
Now in its fifth year, ACM FAccT is considered the most representative international conference in the field of AI ethics. The AI Index Report, presented by Stanford University this year, analyzed the trend in global AI ethics research based on the adoption rate at the conference. This year’s ACM FAccT was held both online and offline, June 21–24.
On the first day of the conference, NAVER held a workshop(CRAFT) with the title of ‘Fairness, Accountability and Transparency in Hyperscale Language Models.’ Domestic and foreign authorities on the AI ethics field participated to discuss ethical issues that may arise in the process of developing a hyperscale language model or applying it to services for actual users, and explore solutions that can solve these problems.
During the workshop, experts from the field of AI ethics gathered to discuss the present, future, and the ethical issues of hyperscale AI. At the first session of the workshop, Ha Jung-woo, head of NAVER AI Lab, introduced the purpose of the event. Also, Professor Ko Hak-soo of Seoul National University (SNU), Professor Cho Kyung-hyun of New York University and Margaret Mitchell, the former head of Google AI Ethics Research Team and the founder of Ethical AI, spoke on the subject of legal and ethical issues of hyperscale AI and the fairness and transparency related to it.
[Photo] Professor Ko Hak-soo(left) of SNU and Margaret Mitchell(right), the founder of Ethical AI is giving a presentation
Experts from various fields related to AI ethics, including professor Cha Mee-young of KAIST, Professor Ko Hak-soo of Seoul National University, Professor Alice H. Oh of KAIST and Professor Park Sang-Cheol of Seoul National University, joined the panel discussion. Industry officials, including Deep Ganguli of AI startup Anthrophic, which was founded by members from Open AI ; Ryu Hwi-jung, team leader of KT; and Sung Nako, Executive Officer of NAVER CLOVA, who oversees the development of HyperCLOVA; also participated in the event and shared their insights. The event was hosted by Professor Park Sang-cheol of SNU and Dr. Lee Hwa-ran from NAVER CLOVA.
On June 22, Chun Sang-hyuk, leader of NAVER AI Lab ML research team, along with Professor Song Kyung-woo of University of Seoul and Jung Yong-han, Ph.D student at Purdue University, conducted a session called ‘Tutorial on Shortcut Learning in Machine Learning: Challenges, Analysis, Solutions.’ The session analyzed the problem caused by a tendency of AI models to prefer easy methods in data learning, and discussed possible solutions.
NAVER is actively leading discussions on AI ethics, while accelerating the development and application of large-scale AI models through Korea’s first hyperscale AI HyperCLOVA. The company published ‘AI Ethics Rules’ in February 2021 through collaboration with SNU AI Policy Initiative (SAPI) for three years and published ‘NAVER-SAPI AI Report’ in November. NAVER also launched an in-house communication channel in the form of a mailing group that can discuss issues related to AI ethics rules when conducting projects or developing services, and plans to publish a Progress Report on the operation of AI ethics rules. A program for startups to share NAVER’s experience in designing and implementing AI ethics rules is expected to be launched, and NAVER is working with the University of Tübingen in Germany to study safe and reliable AI.
“The ACM FAccT workshop is one of the examples of NAVER securing global leadership in the field of hyperscale AI ethics. As AI ethics is a matter of creating social consensus with technical efforts, NAVER will lead and seek ways to develop AI for humanity with various experts,” said Ha Jung-woo, head of NAVER AI Lab.
END