ITRI TODAY Fall issue 2022

Taiwan’s First Online Workshop on Trustworthy AI

ITRI held the 2022 International Workshop on Trustworthy AI online in August, which saw over 500 experts join from around the globe at the first-ever conference on this theme in Taiwan. Topics such as AI risk management and trustworthy AI systems were discussed to explore closely the issues of fairness, transparency, security, and reliability of AI-based decision making.

Minister without Portfolio Audrey Tang lauded the seminar as a good start in promoting the standardization of trustworthy AI and in accelerating the development of digital technology in Taiwan. She pointed out that AI can be understood as Assistive Intelligence, rather than something that completely replaces humans. “It’s like the complementary relationship between the AI robot and humans in the Japanese manga Doraemon,” she said. Taiwan is an important partner in the global supply chain, Tang noted, adding that in the field of ICT, it has formed a solid ecosystem with the US supply chain. The establishment of a trusted AI framework, with its introduction into products and exports to the world, will create a new peak for Taiwan’s industry, while helping Taiwan move closer towards its vision of Smart Taiwan 2030, which emphasizes innovation, inclusion, and sustainability.

ITRI Executive Vice President Pei-Zen Chang remarked that the popularization of AI applications has sparked discussions on ethical issues, including data bias, privacy protection and security, system transparency, and interpretability. For instance, Portable Edge AI-DR, a technology for which ITRI won an Edison Award this year, employs AI analysis to identify the location of disease and its severity in just three seconds, assisting doctors in reaching rapid diagnosis. If this data can be provided to and exchanged among different hospitals while protecting patient privacy and ensuring security, this type of federated learning will then push AI medical care in Taiwan into a new realm.

IDB Director General Jang-Hwa Leu emphasized the importance of AI risk management or supervision. Taiwan intends to cooperate with national laboratories such as the National Institute of Standards and Technology (NIST) in the US and Laboratoire national de métrologie et d'essais (LNE) in France. This will link relevant technical specifications or standards across Taiwan, the US, and the European Union. It will also assist Taiwan’s AI products in meeting common standards and requirements on a global basis, enabling Taiwan to become an internationally trusted partner for AI, he stated.

Also invited to the workshop was Elham Tabassi, Information Technology Laboratory Chief of Staff at the NIST, who shared the NIST's AI risk management tools that can make AI inference as fair as possible based on trustworthy AI models. Meanwhile, Jeannette M. Wing, Executive Vice President for Research at Columbia University, expressed her belief that the trustworthiness of AI systems can be increased thanks to the expansion of trusted computing architecture.

Others attending the event included Andrew Shikiar, Executive Director & Chief Marketing Officer at the FIDO Alliance; Taiwan AI Labs Founder Ethan Tu; Jane Yung-Jen Hsu, Professor of Computer Science & Information Engineering at National Taiwan University; Hon Hai Research Institute CEO Wei-Bin Lee; National Science and Technology Council Deputy Executive Secretary Yuh-Jye Lee; Egis Technology Inc. Principal Engineer Shang-Wei Chou; Tzeng-Yow Lin, Chief Executive of the National Measurement Laboratory and General Director of ITRI’s Center for Measurement Standards; and Kai-Lung Hua, Deputy General Director of ITRI’s Information and Communications Research Laboratories. These experts shared developmental trends and experience in trustworthy AI and held discussions on a technology roadmap for Taiwan.