Why Japan lags behind in Generative AI and LLM development

Japan is known for its futuristic technology. But the country is lagging behind in the race for generative AI and is trying to develop its own large language models.

Mr.cole_photographer | moment | Getty Images

Countries are racing to develop their own generative artificial intelligence algorithms, but high-tech Japan is already lagging behind.

Generative AI has been the trending topic in the technology space since OpenAI caused a stir with its chatbot ChatGPT. Breakthroughs in generative AI have the potential to increase global GDP by 7%, or nearly $7 trillion, over the next decade, according to a study by Goldman Sachs.

The key to generative AI development are large language models underlying ChatGPT and others Baidus Ernie Bot that can process huge datasets to generate text and other content. But Japan is currently lagging behind the US, China and the EU in developing these algorithms, said Noriyuki Kojima, co-founder of Japanese LLM startup Kotoba Technology.

Chinese organizations, including tech giants Alibaba and Tencent, have launched at least 79 LLMs domestically in the past three years, Reuters reported in May, citing research by a consortium of government institutes. US companies like OpenAI, Microsoft, Google And Meta are playing an important role in driving the country’s LLM advances, Kojima said.

Japan lags behind in generative AI

However, Japan is lagging behind the US, China and Europe in terms of the scale and speed of its LLM development.

“Japan’s lag behind in generative AI is largely due to its comparative deficiencies in deep learning and broader software development,” Kojima said.

Deep learning requires a “robust community of software engineers” to develop the necessary infrastructure and applications, Kojima added. However, according to the Ministry of Economy, Trade and Industry, Japan will face a deficit of 789,000 software engineers by 2030. The nation is now ranked 28thth according to the IMD World Digital Competitiveness Ranking of 63 countries in terms of technological knowledge.

Japan also faces hardware challenges as LLMs need to be trained using AI supercomputers IBM’s Vela and Microsoft’s Azure hosted system. But no private company in Japan has its own “world-class machine” with these capabilities, Nikkei Asia reported.

Government-controlled supercomputers like Fugaku are therefore “key” to Japan’s pursuit of LLMs, Kojima explained.

“Access to such large supercomputers is the backbone of LLM development as it has traditionally been the biggest bottleneck in the process,” he said.

How Japan’s supercomputers can help

Tokyo Institute of Technology and Tohoku University plan to use Fugaku to develop LLMs based primarily on Japanese data in collaboration with supercomputer developers Fujitsu and Riken, Fujitsu announced in May.

The organizations plan to publish their research results in 2024 to help other Japanese researchers and engineers develop LLMs, Fujitsu added.

Japan has

The Japanese government will also invest 6.8 billion yen ($48.2 million), about half the total cost, to build a new supercomputer in Hokkaido, which is expected to start operating as early as next year, Nikkei Asia reported. The supercomputer will specialize in LLM education to boost Japan’s development of generative AI, Nikkei Asia said.

In April, Japanese Prime Minister Fumio Kishida said the country supports the industrial use of generative AI technology. Kishida’s comments followed his meeting with OpenAI CEO Sam Altman, who said the company plans to open an office in Japan.

Japanese companies are pursuing generative AI

Big tech companies have also joined the fight to boost Japan’s standing in generative AI. In June, SoftBanks According to local media, Mobile Arm plans to develop its own generative AI platform. This was underscored by SoftBank CEO Masayoshi Son’s announcement that the investment firm plans to shift from “defense mode” to “offensive mode” and increase its focus on AI.

“We would like that [in] “We take a leadership position in the AI ​​revolution,” Son said during a shareholders’ meeting.

SoftBank Group sold its 85 percent stake in SB Energy to Toyota Tsusho in April and recently agreed to sell its 90 percent stake in US investment manager Fortress Investment Group, Nikkei Asia reported. By cutting back on those other investments, SoftBank can free up cash and focus largely on AI through its venture capital investment unit Vision Fund.

SoftBank-owned chip design company Arm will also seek a US IPO later this year. “It will be by far the largest IPO the world has ever seen,” said Amir Anvarzadeh, Japanese equity market strategist at Asymmetric Advisors.

The IPO will provide a substantial sum to boost funds at SoftBank, which posted a record loss of 4.3 trillion yen at Vision Fund for the fiscal year ended March 31.

Arm initially wanted to raise between $8 billion and $10 billion. But as demand for semiconductor chips is “through the roof,” Anvarzadeh suggested Arm could raise as much as $50 billion to $60 billion — or “85% of SoftBank’s market cap.”

He said SoftBank’s stock price is likely to rise, although that’s no guarantee of the success of its AI efforts.

“Basically, I don’t think SoftBank will change Japan’s landscape…they are not a savior of Japan’s AI,” he said.

SoftBank CEO Masayoshi Son says the giant is ready to switch to attack mode

Japanese telecommunications company NTT also announced plans to develop its own LLM this fiscal year, with the goal of creating an “easy and efficient” service for businesses. NTT announced that the company will invest 8 trillion yen in growth areas like data centers and AI over the next five years, a 50% increase from previous investment levels.

Local media reported on the digital advertising company CyberAgent released an LLM in May enabling companies to build AI chatbot tools. The company said it’s one of the few “models that specialize in Japanese language and culture.”

While Japan has yet to catch up in the field of generative AI, Japan is taking the first step with these private sector efforts. Once a “robust infrastructure” is in place, remaining technical challenges should be “significantly mitigated” through the use of open-source software and data from previous pioneers, Kojima said. Bloom, Falcon, and RedPajama are all open-source LLMs trained on vast amounts of data that can be downloaded and explored.

However, companies venturing into this space should expect the competition to stretch out over a “relatively longer time frame,” Kojima said. The development of LLMs requires significant capital investment and a highly skilled workforce in natural language processing and high-performance computing, he explained.

“SoftBank and NTT’s participation in this competition will not change the AI ​​landscape in the short term.”

AI regulation in Japan

The increasing involvement of Japanese tech companies in generative AI development goes hand in hand with a positive attitude towards AI adoption in other sectors. According to a survey by Teikoku Databank, over 60% of companies in Japan are positive about using generative AI in their operations, while 9.1% are already doing so.

hitachi has set up a generative AI center to encourage employees to use the technology safely and effectively, it said in May. With the expertise of data scientists, AI researchers and relevant specialists, the center will formulate guidelines to mitigate the risks of generative AI, the conglomerate said.

Japan will even consider the government’s adoption of AI technology like ChatGPT, provided cybersecurity and privacy concerns are addressed, Chief Cabinet Secretary Hirokazu Matsuno said.

As Japan becomes more open to using generative AI, the government should formulate and enable soft policies for its use while assessing the need for tough regulation based on specific risks, said Hiroki Habuka, a research professor at Kyoto University’s Graduate School of Law.

“Without clearer guidance on what actions companies should take when deploying generative AI, practices can become fragmented,” the professor said.

Leave a Comment