Meiji Era and AI: History Lessons for a Global Enlightenment
How Japan's First Industrial Revolution Can Inform Today's AI Revolution
What if history could give us clues for how we can rapidly adopt AI in a way that benefits society? What if this could be the start of another enlightenment? Not just another regional enlightenment, like the European or Chinese enlightenments, but the first global enlightenment.
Whilst the debates on AI risks are needed, we risk getting so caught up in doomsday discussions that we forget to create a compelling vision for an AI future. Not a utopian sci-fi fantasy but a vision grounded in historical precedents.
As someone that works in technology transformations, I’m fascinated by Japan’s Meiji Restoration. It was the first Asian industrial revolution in Asia. It took place at an unprecedented pace. And, of most interest to me as a tech leader, it was a period of significant tech disruption that didn’t result in unemployment or social issues. As a result, I believe there is much that today's data, tech, business and policy leaders can learn from it.
Before the Meiji restoration, Japan had gone through almost two and half centuries of isolationism. Its rulers, the Tokugawa shogunate, had issued edicts to end virtually all foreign trade and contact. As intended, this protected Japanese society from foreign influence. However, it also cut Japan off from the industrial revolution steaming ahead elsewhere.
This all ended abruptly in 1853 when American steamships, painted black to give them the appearance of floating fortresses, powered their way into Tokyo harbour. Known as the Perry expedition, this trip gave Japanese society a sharp reality check on the pace of technology.
Sound familiar? Whilst there hasn't been a period of deliberate isolationism from AI, AI has progressed for decades without most of society taking much notice. It is only now that Microsoft and OpenAI have made the technology widely available that the public debate has started.
Most of us have yet to get our heads around what today's AI models are capable of, let alone what is coming soon. As Geoffrey Hinton, "The Godfather of AI", put it, "it is like aliens have landed on our planet, and we haven't quite realised it yet because they speak very good English".
Metal steamships were no less alien to a pre-industrial society than sophisticated AI is to today's. How Japan responded to this threat was quite remarkable.
The Perry expedition achieved its aim of forcing Japan to open up to American trade. But the display of power and technology also triggered other changes. Fifteen years after the steamships appeared, a group of young samurai, inspired by concepts of liberalism and frustrated with the pace of change, overthrew the Tokugawa Shogunate and started a series of reforms to modernise Japan and ensure its survival. So began the Meiji era.
Forty years into the Meiji Restoration, Japan was an emerging global power with a per capita GDP equal to Greece's and soon to overtake nations such as Spain. And that isn't even the impressive part. It managed this transition whilst bringing its population with it. During the Meiji period, unemployment levels remained low and average wages for unskilled labour doubled.
Contrast this with the first industrial revolution that took place in the UK. Ned Ludd famously reacted to automation, declining wages and rising unemployment by destroying a stocking frame. In doing so, he kicked off a series of protests against technology and automation that had to be put down by the army and forever left us with the word Luddite1.
Historians reading this may be thinking that Japan's story isn't only one of hope. The Meiji Restoration also contributed to a rise in militarism. Ultimately, just 88 years after American aggression appeared in Tokyo Harbour, Japanese aggression appeared in Pearl Harbour. New technologies can lead to both rising living standards and races for power, as American and Chinese battles for silicon supremacy are showing today.
Whilst there are lessons to learn here, I'd like to focus this article on the positive lessons from the Meiji restoration.
Widespread, inclusive education played a huge role in the success of the Meiji Restoration. Japan didn’t just support people at the start of their lives by overhauling its education system with new schools, universities, libraries, etc. They also engaged with and retrained people already in work with vocational schools and by running expositions and distributing books, papers and magazines to inspire people from all backgrounds and stages of life.
The story of Sakichi Toyoda2 best illustrates the opportunities this created for people. The son of a farmer and carpenter in a poor village, he eagerly read newspapers and magazines and was inspired to form a self-learning group to look for ways to contribute to society. Whilst that didn't deliver immediate results when he read about patents, he was inspired to invent. Rather than destroying looms, like the Luddites, he invented the Toyoda wooden handloom and filed his first patent. Many inventions and patents later, he went on to found the Toyota Industries Corporation.
Let's imagine what Meiji style reforms could look like today. The textbook could become extinct if LLMs become more accurate and provide references. Rather than seeing this as a threat, we should celebrate this. Textbooks are not affordable for everyone and cannot be customised for individual learners. LLMs are already multilingual and could provide the highest standard of customised learning to any student with internet access.
The role of the teacher could shift to that of a facilitator for the students learning. Rather than providing answers, they would ask questions to help their students to think critically and come to their own conclusions. This isn't a new or radical idea. This is the Socratic Method that was practised in ancient Greece.
More recent experiments have shown the potential of this approach to learning in today's era. In Sugata Mitra's inspirational TED talk, he showed how groups of children guided to self-learn got the same results on tests as control groups with trained teachers.
I’ve also seen first-hand the potential for people to learn transformative new skills with little more support than being put into learning groups and directed towards materials. In 2018 I led a data champions programme that aimed to give 200 volunteers the skills to move from doing data work on their PC to working on the cloud. The results were remarkable.
A personal career highlight was when one champion described how they’d been in the same role, doing roughly the same data tasks, with no promotion prospects for more than five years. Less than a year into the programme, they had automated part of their role. Then with the extra time, they did analytics which added value, earned them a promotion and started discussions on where they could develop next.
So what might Meiji style reforms look like today? It might look like personalised education for all and a return to the role of the teacher as a mentor and coach. It might enable large sections of society to retrain, automate their role and define their next role before someone else automates it for them and makes them redundant. It might inspire a new generation of Sakichi Toyoda who will build the organisations for the next century.
To do this will require much more than champion programmes. Learning again from the Meiji era, governments and companies will need to reform education and invest in training and retraining. We will also need campaigns to get everyone excited about how AI could help them and to give them the confidence to learn new skills.
The Meiji era ended in 1912, on the emperor’s passing. The name Meiji (明治) means "enlightened rule". It was a period not unlike the European Enlightenment. Traditional thinking was challenged. Reasoning and the scientific method improved lives and the world.
We now have the opportunity for another enlightenment. With the help of LLMs multilingual capabilities, this could be the first global enlightenment. One where ideas and debates progress across language barriers. Where social media platforms start to broaden, not narrow, our horizons by auto-translating and recommending content from other cultures. In biblical terms, we can rebuild the Tower of Babel itself!
Yes, AI poses risks that we must debate, but let's also create compelling visions for a future with AI. Let's link those visions to historical precedents so that they are grounded in reality. Let's make this the best era to be alive!
Luddite has become a term for a person opposed to new technology or ways of working.
Love the image of the Socratic Method being implemented in schools. Overall my opinion is that AI should make us more, not less, human. Facilitating critical and independent thinking and coming to our own conclusions for sure would make us use our unique human capabilities better.
Great post, especially in light of articles like the one in today's New Yorker, called "Can We Stop Runaway AI?" Your Socratic method point reminded me of the new Khanmigo app from Sal Khan. It guides students in learning but doesn't reveal any answers: https://www.khanacademy.org/khan-labs