Reflections on How AI Reshapes Human Society

Recently, I’ve seen many articles comparing the AI revolution to the Industrial Revolution in human history. However, after careful consideration, I believe that the AI revolution is fundamentally different from the Industrial Revolution and will have a more profound impact on the structure of human society.

Those comparisons to the Industrial Revolution liken AI to the steam engine, electricity, and assembly lines of the past, suggesting that it will greatly increase the overall productivity of society, just like the Industrial Revolution did. This statement is not entirely incorrect, but I feel that it doesn’t capture the essence of AI.

In my opinion, the Industrial Revolution was more like a universal enhancement of human capabilities. For example, in the past, a person could only make one thing by hand in a day, but with the help of machines and assembly lines, they could make a hundred things in a day. While there are still differences in individual abilities, the improvement brought by machines is more like a “universal bonus.” Your original personal ability is important, but it’s not the deciding factor.

But AI is different.

AI is more like an amplifier of abilities, rather than a uniform mechanical system.

For instance, a person with an original ability of 1 may become 10 with the help of AI. Another person with an original ability of 100 may become 1000 with AI. On the surface, it seems like both people have improved by a factor of 10, which appears fair. However, in reality, the gap between them has expanded from 99 to 990.

Another example is that in ancient times, swordsmen fought with swords, with varying levels of skill. The Industrial Revolution was like giving everyone guns, making it difficult for skilled swordsmen to defeat enemies with guns, even if they had mastered their sword-fighting skills. However, it’s possible that a weak soldier could become a skilled gunman. This is the social redivision of labor brought about by the Industrial Revolution, where swordsmen are phased out, and everyone is forced to become a gunman.

However, the AI era is more like giving each swordsman an AI-enchanted sword that can amplify their original attributes by different proportions, up to 1000 times. In this scenario, the originally strong swordsman will become invincible, while the weak swordsman, although stronger, will become even weaker in comparison.

This is where I think the AI era differs from the Industrial Revolution: it’s not simply lifting everyone up together, but rather amplifying the existing differences between individuals.

Moreover, the more capable a person is, the more they can truly utilize AI. AI is not something that automatically generates value when turned on; it requires you to know what you want to do, to ask questions, to judge, to filter, to integrate, and to have resources, time, and space for trial and error. A strong person using AI is not just “saving some effort” but is instead entering a state of explosive growth. On the other hand, a person with a weaker foundation, even if they can use AI, often only stays at the level of “being able to use it” rather than “using it extremely well.”

Therefore, I increasingly feel that the AI era may make society even more unfair.

Especially when I consider this issue in the context of reality, it becomes even more apparent.

The rich, the strong, and large companies already possess more resources. They not only can use AI but also have access to more advanced AI (such as the mysterious Mythos): better models, longer context, higher call quotas, more private data, more automated processes, and even the ability to build their own complete work systems around AI. For them, AI is not just a single tool but an entire set of amplifiers.

On the other hand, ordinary people and the poor may only have access to free AI or AI of average quality. Due to limitations in time, devices, education background, and other factors, they may not be able to truly integrate AI into their work and life. As a result, the gap between them is not slowly widening but is instead being rapidly expanded.

I think that if we really want to find an analogy for the impact of AI on society, we should compare it to the biological competition model in nature rather than the Industrial Revolution.

For example, I’ve been thinking about a very intuitive example recently: suppose we have an area with both trees and grass, and we fertilize them together, causing their production to increase by 10 times. What would be the result?

Many people might instinctively think that everyone would grow better. However, the reality is not necessarily so.

Because trees are originally taller, they get sunlight first; their roots are also deeper, making it easier for them to absorb water and nutrients; and their larger size allows them to expand more quickly when resources increase. The result is not that both trees and grass thrive simultaneously, but rather that trees grow faster and taller, their canopies becoming denser and gradually shading out the grass. Eventually, this area will be dominated by tall plants, making it harder for smaller plants to survive.

I think AI’s impact on human society is somewhat like this “fertilization.”

Resources are not decreasing, and productivity is not declining. The problem is that new resources and capabilities are not being distributed evenly to everyone but are instead being absorbed more efficiently by individuals and organizations that already have an advantage.

In this way, the future social structure may not be one where everyone becomes stronger together but rather where a few originally strong individuals and organizations use AI to create a huge “canopy advantage.” They gain more information, more wealth, more decision-making power, and stronger organizational capabilities, while many ordinary people, although also accessing AI, only receive marginal improvements and may even be further squeezed out by the efficiency explosion at the top.

Of course, I don’t think things will be as simple as “the strong taking everything.” AI will indeed lower many thresholds, allowing people who originally lacked professional capabilities to do things they couldn’t do before. Some old skill barriers will be broken, and new people may even have opportunities to leap forward through AI.

However, overall, I still believe that the core characteristic of the AI era is not averageization but amplification; not reducing differences but multiplying them exponentially.

So, if we really want to understand how AI will change society, perhaps we shouldn’t just focus on “how much productivity has increased” but rather look at: who is absorbing these improvements? Who can turn AI into a long-term advantage? Who can only stay at the surface level of using AI?

From this perspective, what’s most worth warning about in the AI era is not simply “machines replacing humans” but rather: will the originally strong people become so strong with AI that ordinary people can’t catch up? Will those who are already rich become even richer by amplifying their abilities 100 times with AI? And will those without resources, without access, and without organizational support be left further behind in this process?

This is what I think is truly frightening and worth serious consideration about AI.

The future world described in cyberpunk novels has really arrived.