As the need for legislative reforms arises in response to the introduction of autonomous driving in mobility, ships, and urban air mobility (UAM), discussions on related issues are heating up. The insurance industry anticipates a reduction in human error and accidents involving death or disease due to unmanned and autonomous operations, and expectations for the maturation of artificial intelligence technology are also rising.

The 2022 Artificial Intelligence Legislation Reform Team public seminar was held at the Federation of Korean Industries Conference Center in Yeouido on the 22nd. This seminar, organized by the Korea Intelligent Information Society Agency, was held under the theme of "Responsible Artificial Intelligence" to discuss legislative reforms and improvement measures in response to major issues in artificial intelligence.

At the presentation, experts in the humanities, social sciences, finance, and law gathered to conduct presentations and discussions centered on clarifying the validity of contracts made by artificial intelligence, preparing measures to reform the insurance system for handling AI-related accidents, and ensuring transparency in AI administration.

In particular, among industries utilizing artificial intelligence, autonomous driving technology currently undergoing commercialization and countermeasures for the resulting improvement of insurance systems attracted attention. Professor Ji Gwang-un of Kunsan National University presented research findings on insurance responses to the emergence of autonomous driving and UAM technologies, under the theme of AI accidents and insurance.

In automobile accidents, while manufacturers have been liable for compensation under product liability depending on the cause of negligence, the traditional party responsible is the owner. However, if accidents caused by artificial intelligence errors reach the stage of a fully autonomous driving system, we may enter an era where liability for compensation shifts from the 'owner' to the 'operator'.

Professor Ji Gwang-un mentioned the possibility, stating, “Policies could be designed to either place heavier responsibility on manufacturers or to encourage system operators to sign up for insurance.” He stated, “The importance of cybersecurity will rise alongside hacking,” suggesting that the issue of establishing safety standards to respond to cyber threats is also being discussed globally.

In the field of autonomous vehicles, Japan and Germany have reached Level 3, and recently in Korea, Hyundai Motor Company has announced the launch of a Level 3 autonomous vehicle in January 2023, demonstrating that autonomous driving technology is maturing rapidly.

Professor Ji explains that the liability system and responsible parties do not change significantly up to Level 3 autonomous driving. Accordingly, he predicted that in the event of an autonomous driving accident, the existing insurance system will be able to handle it, and when Level 4 fully autonomous driving is reached, liability will shift to manufacturers and operators, leading to a paradigm shift from private-sector-centered insurance to operator-centered insurance.

During the discussion session, Kang Dong-hwa, a manager at Korea Shipowners Mutual Insurance, explained, “Even in autonomous vessels, up to Level 3, the concept of an unmanned vessel is maintained, with a land-based control master controlling the ship.” He predicted that at this level, land-based crew members, such as control masters who remotely control the vessel, will be the primary responsible parties, whereas for Level 4 fully autonomous vessels operated by a system, it will be necessary to monitor the progress of relevant discussions, such as international conventions.

Director Kang said, “Among the causes of maritime accidents worldwide, accidents caused by injuries and deaths of crew members account for 40%, and human error, such as crew operational mistakes, accounts for 40%.” He mentioned the industry's expectations regarding autonomous ship operation, stating, "Since accidents still occur even with crew members on board, the relevant industry anticipates that insurance claims related to death and disease will decrease once technology beyond unmanned vessels is introduced."

In addition, a session on securing transparency in AI administration led by Professor Hyang-Won Kwon of Ajou University followed, emphasizing the need to be wary of data bias in deep learning and machine learning to ensure transparency in the use of AI in administration. Professor Kwon discussed ethical issues regarding artificial intelligence, such as copyright problems, algorithms reflecting the specific intentions of designers, and bias in machine learning.