Silicon Valley’s “move fast and break things” mantra propelled tech innovation for the Internet age. In the era of artificial intelligence (AI), it should take a leaf out of Japan’s playbook and slow down.
A rush to deploy AI tools to the public has resulted in embarrassing blunders, from an AI-powered Google search feature recently recommending glue on pizza, to consequences that can impact real people’s livelihoods, like the technology behind OpenAI’s ChatGPT showing signs of racial bias when ranking job applicants, as a Bloomberg analysis found.
It has also led to tech companies consuming enormous amounts of energy to power AI. The International Energy Agency estimates that the total electricity consumption for data centers across the globe would be about the same as the power demand of Japan in 2026. Other forecasters say that by 2030, these centers are on course to use more energy than India, the world’s most populous country. Large language models, the technology underpinning the latest crop of generative AI tools, require gargantuan troves of data and training them takes immense amounts of computing power and energy.
Illustration: Yusha
As the tech continues to develop, many AI firms think the key to growth is to make these large language models even larger. Some US tech titans including Microsoft Corp cofounder Bill Gates and OpenAI chief executive officer Sam Altman are even backing nuclear energy firms to help power AI data centers. However, there are other options beyond rushing to fire up new nuclear reactors to train AI models.
Tokyo-based startup Sakana AI — which Nikkei Asia reported on June 15 would become the fastest-ever Japanese company to achieve unicorn status — has taken a different approach. When it comes to the most consequential technology of our time, they are playing the long game.
“Sakana,” which means “fish” in Japanese, was created last year by David Ha, Ren Ito and Llion Jones with the goal of applying nature-inspired ideas such as evolution and collective intelligence to create AI models, and not gobbling up as much energy as possible to train a single AI model. Jones, notably, coauthored the seminal 2017 research paper while at Google that underpins the technology of today’s most popular AI products, including ChatGPT, and spurred much of the current boom.
Ha told me that much of the issues facing AI stem from companies prematurely deploying the technology in search of short-term profits, while disregarding the high energy toll. While Sakana’s technology could be used to create a virtual girlfriend chatbot now, they are more focused on researching ways to apply AI to solve real-world issues. This is in line with other Japan-based AI companies the West might have never heard of, such as Preferred Networks, which is trying to develop more energy-efficient AI chips.
Silicon Valley needs to start playing the more long game, too. The world does not need more useless chatbots that generate bad poetry — or perpetuate racist biases at scale — while consuming unsustainable swaths of resources. This way of thinking could have also spared some of the embarrassing headlines that have turned much of the public against embracing AI for good.
Sakana’s most recent research is looking at new ways of producing AI models that do not require extensive amounts of energy, and it has longer-term ambitions of helping Japan at a national scale come up with real solutions to some of its most pressing issues, including a looming labor crunch. Their mission has received support from the Japanese government.
In Japan, depopulation and a shrinking workforce have led to a public and private sector ecosystem that seeks to embrace the emerging technology’s possibilities with curiosity and hope instead of despair. There is a sense of urgency around the idea that these tools can help solve labor shortages — with less of the existential fears that they would end up replacing human workers. Microsoft Japan president Miki Tsusaka touted generative AI’s power to accelerate growth in a nation with an aging population in an interview with Bloomberg TV last week. Her comments come on the heels of Microsoft announcing it would spend about US$2.9 billion, its biggest-ever investment in Japan, to increase cloud computing and AI infrastructure.
Being a close US ally has also helped Japan in the current boom — it has welcomed a spate of US tech investment that might have previously been funneled toward China.
Along with Microsoft, OpenAI opened its first Asia office in Tokyo this spring, while Oracle Corp pledged to put more than US$8 billion over the next 10 years into cloud computing and AI.
Risk-averse Japan lost its place as a global tech leader during the past few decades, but the AI boom, still in its infancy, is an entirely different beast. This time around, its cautious and collective approach to the technology means Tokyo is uniquely positioned to come out ahead when it comes to getting AI right. The rest of the world might soon find itself playing catch-up.
Catherine Thorbecke is a Bloomberg Opinion columnist covering Asia tech. Previously she was a tech reporter at CNN and ABC News. This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.
Elbridge Colby, America’s Under Secretary of Defense for Policy, is the most influential voice on defense strategy in the Second Trump Administration. For insight into his thinking, one could do no better than read his thoughts on the defense of Taiwan which he gathered in a book he wrote in 2021. The Strategy of Denial, is his contemplation of China’s rising hegemony in Asia and on how to deter China from invading Taiwan. Allowing China to absorb Taiwan, he wrote, would open the entire Indo-Pacific region to Chinese preeminence and result in a power transition that would place America’s prosperity
When Democratic Progressive Party (DPP) caucus whip Ker Chien-ming (柯建銘) first suggested a mass recall of Chinese Nationalist Party (KMT) legislators, the Taipei Times called the idea “not only absurd, but also deeply undemocratic” (“Lai’s speech and legislative chaos,” Jan. 6, page 8). In a subsequent editorial (“Recall chaos plays into KMT hands,” Jan. 9, page 8), the paper wrote that his suggestion was not a solution, and that if it failed, it would exacerbate the enmity between the parties and lead to a cascade of revenge recalls. The danger came from having the DPP orchestrate a mass recall. As it transpired,
A few weeks ago in Kaohsiung, tech mogul turned political pundit Robert Tsao (曹興誠) joined Western Washington University professor Chen Shih-fen (陳時奮) for a public forum in support of Taiwan’s recall campaign. Kaohsiung, already the most Taiwanese independence-minded city in Taiwan, was not in need of a recall. So Chen took a different approach: He made the case that unification with China would be too expensive to work. The argument was unusual. Most of the time, we hear that Taiwan should remain free out of respect for democracy and self-determination, but cost? That is not part of the usual script, and
All 24 Chinese Nationalist Party (KMT) lawmakers and suspended Hsinchu Mayor Ann Kao (高虹安), formerly of the Taiwan People’s Party (TPP), survived recall elections against them on Saturday, in a massive loss to the unprecedented mass recall movement, as well as to the ruling Democratic Progressive Party (DPP) that backed it. The outcome has surprised many, as most analysts expected that at least a few legislators would be ousted. Over the past few months, dedicated and passionate civic groups gathered more than 1 million signatures to recall KMT lawmakers, an extraordinary achievement that many believed would be enough to remove at
RSS