AI Adoption Narrative Moving Quickly in Japan

How Far We’ve Come

There’s a well-trodden adage about Japan which posits that the country goes beyond what is typically thought of as reasonable deliberation on the adoption of new policies, but that when it does finally choose to implement a new resolution, it does so with incredible speed and efficiency. 

Honestly? There are some historic examples of this (Meiji, anyone?), but these days it feels to us that it’s more the exception than the rule. Just because Japan has implemented sweeping societal changes in the blink of an eye doesn’t mean it’s prone to doing it. In fact, particularly in the IT sector, there’s contemporary evidence which points to an unfortunate misalignment between accepted domestic practice and global market realities. 

It feels like just last week that we were reporting on Japan’s sluggish and stubborn approach to engaging with (different from adopting) AI platforms. That narrative now appears to be veering rapidly from the established course and into a new direction of curiosity and experimentation, which we welcome quite openly.

OpenAI CEO Sam Altman Visits Japan

The first notable development is also likely the most powerful one. Where there was skepticism amongst Japanese organizations, OpenAI CEO Sam Altman saw promise, enough to drop in to say hi to Prime Minister Kishida. Reuters reported on this on April 10th, and you can read a summary below:

OpenAI is considering opening an office in Japan and expanding its services in the country following a meeting between Prime Minister Fumio Kishida and OpenAI CEO Sam Altman.

During the meeting, they discussed the technological progress, benefits, and risks of AI, including privacy and copyright infringement.

Japan will assess the possibility of introducing AI-powered technology such as OpenAI's ChatGPT chatbot, subject to assessing the benefits and risks. The chatbot has raised privacy concerns and caused Italy to temporarily ban it.

Altman hopes to build something great for Japanese people, make the models better for Japanese language and culture, and expressed hope that Japan will play a role in adopting AI and rule-making. The government is considering government adoption of artificial intelligence technology such as ChatGPT if privacy and cybersecurity issues are resolved. Japan will continue evaluating the possibilities of introducing AI to reduce government workers' workload after assessing how to respond to concerns such as data breaches.

This was a surprise to us - as Shinka Labs CTO Aaron noted on a Linkedin post - and almost entirely in the good sense. That the person in the public AI spotlight honed in on Japan to begin the AI World Tour is of considerable importance, though we can only guess the why - Perhaps Altman saw Japan dragging its feet and wanted to give them a push?

Japan’s Financial Institutions Opening to AI

The next big development is that, in a abrupt about-face, the major financial institutions of Japan including Daiwa, Mizuho, and MUFJ have decided to at least discuss specific implementations of AI into their businesses. Bloomberg published a story on this on April 18th, which is summarized below:

Daiwa Securities Group is planning to adopt OpenAI's ChatGPT technology for all of its 9,000 domestic employees. The Japanese firm said the technology, which can help staff gather information in English as well as save time and cut costs, has "limitless potential".

The move follows Mitsubishi UFJ Financial Group's reported plans to use a chatbot for drafting approval requests and internal inquiries, and Mizuho Financial Group's intention to launch an AI tool for internal use.

Other financial services firms, such as Sumitomo Mitsui Financial Group, are reportedly trialling AI chatbots. Daiwa said it would explore new ways to use the technology.

Sam Altman, CEO of OpenAI, met with Japan's Prime Minister Fumio Kishida in Tokyo earlier this month to discuss the potential of the technology, as well as how to mitigate the downsides. While some Japanese companies are embracing AI chatbots, Wall Street banks, including Bank of America, Citigroup, Deutsche Bank, Goldman Sachs and Wells Fargo, have banned the use of ChatGPT over concerns about its accuracy and security.

While we do consider this a positive development, and applaud the effort to engage with these nascent technologies by some of Japan’s most stubborn organizations, it doesn’t come without concern. Financial institutions are, by their nature, pretty conservative for very good reasons - if they screw up, the economy crashes. 

So in principal, we can’t really dock any points from Japan Inc. for being cautious. But we do also get a sense that they’re cautious because they know their systems are fragile, and that they aren’t developing new ones, thus perpetuating the fragility and need for caution, and even that caution isn’t enough because, for example, Mizuho ATMs had multiple outages last year.

Which is to say, they might have reason to fear AI - they might not have the chops to handle it.

Japan Hosts G7, Calls for AI Governance

The last piece of news we’ll report on here is perhaps the least exciting: Japan is hosting the G7 Conference (you know, that intergovernmental political forum consisting of Canada, France, Germany, Italy, Japan, the United Kingdom and the United States) this year in Hiroshima, and has taken the liberty of announcing their intent to work with other members in formulating AI governance policy.

Granted, this is a pretty “no duh” sort of development. AI is emerging technology that governments everywhere are unsure of the extent of its power - it’d be weird if they weren’t interested in defining acceptable use boundaries. But where it gets interesting is that it potentially lights a fire under Japan - and the rest of the G7 members’ - collective ass. The Mainichi released a short piece about this on April 20th, but the gist of it is:

The G7 digital ministers will call for accelerated research into generative artificial intelligence systems, such as OpenAI's ChatGPT, at a meeting later this month, according to a draft communique. The ministers aim to formulate an action plan on AI governance during their two-day talks, and the Japanese government is considering discussing generative AI at a G7 summit in May.

ChatGPT can simulate human-like conversations with users and has utility in various fields. However, many countries have been stepping up regulations on its use over concerns that OpenAI illegally gathers vast amounts of personal data from users.

The draft statement also proposes formulating countermeasures to tackle fake news and other forms of disinformation and establishing an international framework for Data Free Flow with Trust, a concept proposed by Japan at the World Economic Forum annual summit in 2019.

Additionally, the G7 will draw up an action plan for constructing secure network infrastructure to support emerging and developing countries, improve the connectivity of undersea cables, and address the blocking or limiting of internet access by some countries.

Knowing that Japan’s also recently released their Web3 roadmap and whitepaper, that Data Free Flow with Trust aspect is particularly interesting. As recently pointed out in an article by The Guardian regarding German artist Boris Eldagsen’s winning of a Sony’s world photography award with an AI-generated image, reality is a mix of truth, hallucination, and non-truth. Being able to distinguish these from one another is really at the crux of the whole world’s trepidation regarding AI, but maybe, just maybe, Japan is suggesting that we… make NFTs of everything?!

Fun Times

All of these stories are great developments. As we’ve said before, making an informed decision about using a new technology - be it a “yes” or a “no” - is one of our major missions; forcing risky tools upon unprepared organizations is not. Japan Inc’s got a pretty good recent record in keeping a distance from anything too uncertain, and indeed they started the Age of AI with that characteristic approach. 

But if the developments over the last week or so are anything to extrapolate from, we might be in for a more interesting ride in the coming days, weeks, and months.