Generative AI Pushing Life Sciences to New Boundaries
On February 22 the Nikkei 225 index finally surpassed the peak it reached at the height of the bubble economy in 1989, making headlines and provoking analysis in Japan and around the world. One of the major drivers of the current buoyancy in global equities is optimism around how the spread of Generative Artificial Intelligence (AI) will lift demand for semiconductors and boost efficiency across much of the economy. Japan’s healthcare sector and the wider life sciences field are ripe for an AI overhaul, a process that has already begun.
It is the ability of this new iteration of AI to generate – text, code, images, audio, synthetic data and more – that sets it apart from previous deep learning models, and is sparking hopes that it could change how business is done as drastically as the internet did. Widespread adoption of generative AI over the coming decade could expand the global economy by 1.5%, adding $7 trillion to GDP, suggests research by Goldman Sachs.
Illustrating the belief that generative AI offers a holy grail, on the same day the new record was set, semiconductor equipment maker Tokyo Electron became Japan’s third most valuable company by market capitalisation. On the way it leapfrogged Sony, a firm with its own sizable semiconductor division.
Native tongues and supercomputers
One of the perceived barriers to full deployment of AI in Japan is that the large language models (LLM) of interfaces such as OpenAI’s ChatGPT and Google’s Gemini (known until recently as Bard) are unable to grasp the subtleties of Japanese. To address this, the government is working with the private sector and academia to spur the creation of native Japanese LLMs.
Supercomputers are crucial to provide the vast amounts of processing power required; Fugaku, developed by Fujitsu and government-funded research institute Riken, is being employed in Japan’s pursuit of LLMs in a partnership with Tokyo Institute of Technology and Tohoku University. The government has also announced it will fund half the costs for a new supercomputer to be built in Hokkaido by cloud provider Sakura Internet. Developed specifically for LLMs, it is set to come online as early as this year and will approximately triple the nation’s generative AI capacity.
The Ministry of Economy, Trade and Industry (METI) is involved in a supercomputer project through the National Institute of Advanced Industrial Science and Technology (AIST). Financing of up 32 billion yen ($226 million) will establish a new research centre to develop cutting-edge supercomputer and quantum technologies, with the computing facilities to be made available via the cloud to domestic companies working on generative AI. METI is also providing funding of 8.4 billion yen ($56 million) to seven entities, including startups and a university, for AI development.
Generating new ideas and treatments
Meanwhile, the education ministry has teamed up with RIKEN on a generative AI programme to enhance medical and scientific discovery. The system will work by analysing research papers and experiments in order to produce new hypotheses, identifying causes of illness and designing new medical and industrial materials. As with the other projects, the intention is to boost both domestic capacity and economic security by reducing the need to rely on tech from overseas.
In the private sector, trading house Mitsui & Co is collaborating with chip giant NVIDIA on a generative AI system designed specifically for drug discovery. The Tokyo-1 project will feature a supercomputer that will be accessible to Japan’s pharma firms to facilitate accelerated drug development. Though home to major pharma corporations and innovative treatments, Japan has lagged behind in some areas of life science and healthcare, including new drug development and approval. This was highlighted by the slow development and rollout of coronavirus vaccines.
Cross-border and underway
Healthcare generative AI tools already deployed in Japan include a platform that summarises preliminary interviews with patients for doctors. Developed in conjunction with doctors by medical startup Ubie, it was launched last summer and is being used by healthcare providers in all of Japan’s 47 prefectures. Ubie debuted its AI-driven symptom checker app in the US in 2022, racking up around 2 million uses last year.
California-based RapidAI made the reverse journey, receiving approval in Japan in late 2023 for its early detection system of strokes from CT scan images. The startup has also announced an innovative hybrid cloud and on-site platform that allows healthcare providers to continue to use AI even in the event of lost access to remote servers.
More to do, more to come
At its best, health AI will deliver personalised and optimised medical treatments to empowered and informed patients. Provision will be efficient, reasonably priced and available wherever needed, even remotely where possible. This transformative tech also holds the promise of alleviating the personnel shortages and budgetary constraints already impacting Japan’s healthcare system.
Nevertheless, challenges remain. Privacy and data protection will become more important than ever, while creating regulations to control potentially wayward AI systems will be a herculean task. Allowing groundbreaking research in medicine and life sciences while at the same time ensuring safety and protection of individual rights is not an easy balance to strike.
Workforce shortfalls are also to be found in IT, with the deficit in software engineers expected to be nearly 800,000 by the end of the decade. Unaddressed, this could severely impact utilisation of generative AI everywhere in Japan. A silver lining to that particular problem will come in the form of abundant opportunities and a strong bargaining position for those with the relevant skills and knowledge to steer the coming together of generative AI and life sciences.
By: Gavin Blair
We are currently hiring for:
AI Data Scientist- Drug Discovery
Commercial IT/DX Project Representative
Data Architect
Global Infrastructure Integration Expert
IT Manager Technical Product Owner
IT Senior Technical Support, Site IT Technical Japan
IT/DX project in R&D department
Manager, Data Management & Statistical Analysis
SAP SD Manager
Senior Application Architect
Senior Data Scientist/Data Scientist, Medical HQ, EOR, Data Science
Senior Infrastructure Architect
Solution Architect
To find out more about hot roles in healthcare and elsewhere please contact us at +81-3-5962-5888 or email us at info@slate.co.jp