Japan is gearing up to revolutionise scientific discovery with the development of a cutting-edge generative AI programme.

In a significant move to boost scientific research and development, Japan’s education ministry (MEXT) has revealed plans to develop a homegrown generative artificial intelligence (AI) programme. The AI will generate medical and scientific hypotheses by learning from research papers and experimental images. This decision comes as a measure to safeguard data and enhance national competitiveness, as reliance on foreign technology could pose risks of information leaks.

The ambitious project, spearheaded by the renowned Riken research institute, will initially focus on medical and material research and later expand into other areas. The estimated cost for developing generative AI for one research domain is approximately US$212 million. The ministry is set to seek funding for the initial development phase in the fiscal 2024 budget.

Japan generative AI


The programme’s timeline spans eight years, with a trial period planned to begin in fiscal 2025, allowing external laboratories and companies to access the technology. The broader research community can expect nationwide availability of the generative AI from fiscal 2031.

The core functionality of the developed generative AI will be identifying disease-causing substances and designing materials for medical and industrial applications. Leveraging Riken’s vast repository of accumulated research data, the AI will be trained on massive amounts of information related to candidate compounds and diagnostic images, making it an ideal tool for research in both medical and material fields.

Furthermore, the newly developed AI technology will aid researchers in preparing papers by analysing past literature. As the technology evolves, researchers will interact with the AI to propose and test novel hypotheses, significantly accelerating the research process.

Early estimates indicate that the adoption of generative AI could reduce the time taken from idea conception to paper publication to less than one-tenth in certain research domains, positioning Japan for heightened international competitiveness.

In Japan, companies such as NTT and SoftBank are actively working on models compatible with the Japanese language. The technology’s potential for enhancing productivity extends beyond scientific research, encompassing areas such as email, document writing, and minute keeping.

To complement this project, the Japanese government will enhance its research infrastructure and the Ministry of Economy, Trade, and Industry has already scheduled the introduction of a new supercomputer in 2024. Additionally, MEXT will augment the computing power of Riken’s Fugaku supercomputer to facilitate generative AI research.

With Japan pioneering the development of its own generative AI and strengthening research infrastructure, the nation is poised to embark on a new era of scientific discovery and innovation.

Featured banner image credit: biospectrumasia.com


Related Articles

Artificial Intelligence, Nanotechnology, and Other Top Tech Growth Sectors to Quadruple by 2025

AI Technology to Streamline Bureaucratic Processes in Japan

Japanese AgriTech Startup Farmship Leverages AI to Reduce Food Loss