Japan is the latest country to greenlight artificial intelligence governance regulation, landing on a statute that supports innovation and development while acknowledging potential risks.
The Act on the Promotion of Research and Development and the Utilization of AI-Related Technologies passed the House of Councilors 28 May after making it through in the House of Representatives in April. The bill makes clear the legislature wants to see AI thrive in the country, stating up front that “artificial intelligence-related technologies are fundamental technologies for the development of Japan’s economy and society.” It calls its structure “basic” and relies on business cooperation and current laws to regulate the technology rather than inventing a new structure.
The bill still needs to be submitted to the Cabinet and approved by the emperor for the law to be promulgated. The bill’s text states Articles 3 and 4, which handles research and development principles to be created by the government, must take effect no more than three months after enactment.
Also Read: CloudSign Expands to Five Municipalities in Fukushima
Japan’s law is another example of how members of the Asia-Pacific region, including Singapore and South Korea, are taking a more relaxed approach to AI governance for now compared to the EU.
The nation has led some international efforts to find common principles for managing AI, including the Hiroshima AI Process, which was a Japanese-led initiative that produced an international framework on life cycle governance.
Thong said the bill arose out of a desire to promote economic development through AI, something the Liberal Democrat Party emphasized in a white paper outlining its position on the technology, alongside public concerns about AI.
Both Japan and South Korea’s AI laws center around government-supported research and development. This could lead to others in the APAC region following suit, according to Future of Privacy Forum Asia-Pacific Managing Director Josh Lee Kok Thong.
“Whether this trend actually plays out, however, will depend on whether such regulatory innovation is in fact seen as driving AI innovation and digital growth in Japan and South Korea,” he said.
The bill has no penalties but gives the government the ability to advise businesses if they are using AI in a harmful way and to provide guidance on how to fix the situation. The government is also empowered to disclose when malicious actors are identified and rely on current laws regarding personal information protection or copyright to pursue enforcement.
This is not unusual in Japan, where administrative penalties are often the last resort, said Thong, pointing to the country’s Act on the Protection of Personal Information as an example.
“Policymakers may thus believe that this should be sufficient to ensure broad alignment with existing AI guidelines published by the Japanese Government and a manageable level of AI risk,” he said.
And governance folks should be mindful that while the bill does not have a rules-enforcement mechanism yet, it empowers the government to develop one down the road, noted APAC GATES Managing Director Seth Hays.
“This is intentional and meant to set guardrails for firms to be responsible and adhere to soft law guidance already available, but not deter investment,” he said.
The bill presses for transparency from operators on how AI is developed and used by operators, noting AI can “lead to criminal use, leakage of personal information, copyright infringement, and other situations that harm the peace of people’s lives and their rights and interests,” if used improperly.
Hays said other effects from the bill will occur downstream, such as the formation of an AI strategy headquarters. A task force headed up by Prime Minister Shigeru Ishiba will create guidelines for businesses and companies to follow.
And friction over some risks, like copyright infringement cases, might arise through social consensus. Hays pointed to the country amending its copyright rules in 2019 to include an exception for data training. Regulators encouraged disputes between creators and technology companies to be resolved through measures like licensing.
“Similar expectations might be found where AI innovation and disruption occur,” he said.
Future legislation may touch on areas where consensus around AI harms have been growing, such as deepfakes and explicit content. The Japan Times reported the Lower House Cabinet Committee attached a provision to the bill calling for stronger protections against these uses.
SOURCE: IAPP
