How Far Could AI Widen the Gap Between People Who Act and People Who Cannot?

Give two people the same image-generating AI, and one may be satisfied with making an image while the other tries to build a system that automatically handles generation, selection, and sales.

They start with the same tool. But one enjoys it in the moment, while the other tries to leave behind something they can use again tomorrow.

This is where I felt something frightening about inequality in the AI era.

Previously, even someone with motivation and ideas might stop because they did not know the technology. AI lowers that barrier. Someone who responds, “If I can do that too, then let's try this next,” may be able to go further than before.

Everyone becomes able to do more. That does not necessarily bring everyone's results closer together.

What makes this difficult is that I, too, sometimes cannot get myself to act. Even when someone explains the usefulness and I understand the reasoning, I may simply think, “Oh, okay,” and leave it there.

So this is more than an upbeat story about people who take action winning. How will society deal with the differences that emerge after people gain access to the same tools?

People Play the Same Game Differently

When discussing this with AI, I used games as an example.

Given the same game, some people thoroughly explore and master it, while others play a little and stop. Providing the same tools does not make everyone's relationship with those tools the same.

What I previously saw in games as a difference in how much people enjoy them could, with AI, become a difference in what they create and how far they can take their work in the real world.

And for the person deeply engaged, it may feel more like play than effort. As they explore, thinking “It can do this too?” and “What else could I try?”, they gain experience both in using the tool and in completing things.

One person tries a little and stops; the other keeps playing. Could that difference grow into a gap in experience useful for work and creation?

I think of differences arising from this engagement with tools as a “gap in taking action.”

What You Build Helps You Act Again

If we look only at the first step, the difference may seem small.

Make one image. Change how you make it a little. Choose something you like.

But thinking, “Could I automate the repetitive part?” changes the picture. Instead of performing the same operation each time, you try to build a reusable system.

Once the system works, you can spend that time reviewing results or trying other improvements. If you connect it to selection and sales, you can also gain experience delivering what you create and receiving people's reactions.

That feedback, and the tools you built, can be used in the next attempt.

The result of acting once reduces the effort required to act next time.

If this cycle works, the difference is no longer simply that one person tried ten times and the other once. The person who tried ten times also has a system that makes the next ten attempts easier.

Automated processes can run without the person operating them every time. Of course, checking and revision remain necessary, and connecting a system to sales does not guarantee that anything will sell. Even so, the conditions for starting the next thing have changed.

A small difference in willingness to try becomes a difference in what can be done next, and that difference supports further action. Could AI accelerate this accumulation? That is the heart of what frightened me.

Supporting Ability and Equalizing Results Are Different

At least for writing tasks, research has found that AI narrowed performance differences between participants.

In a 2023 study, Noy and Zhang assigned writing tasks to 453 college-educated professionals and gave half access to ChatGPT. Average completion time fell by 40%, and output quality increased by 18%. They also reported that performance differences between participants narrowed. Paper published in Science

But becoming more alike in the ability to complete an assigned task differs from choosing your own tasks, repeatedly experimenting, and leaving behind systems you can reuse.

This study did not investigate long-term differences in taking action. My concern is that AI may compensate for weaknesses while other differences emerge in the range and accumulation of activities people initiate themselves.

Sometimes I Cannot Act Either

In the original conversation, my starting point was: “I'm the same—when I can't get myself to act, I just can't.”

I understand the reasoning but do not become interested. That happens to me too.

With AI, trying it myself made me feel, “This is serious.” Fear was closer to my reaction than simple admiration for its usefulness. At the same time, it felt like a toy box.

That fear and enjoyment prompted me to experiment.

But people who hear the same explanation and try the same tool will not necessarily all feel the same way.

For someone eager to keep experimenting, AI expands possibilities. Yet remembering how I feel when I cannot act, I cannot simply say, “You can do it, so just do it.”

As more becomes possible, people may hear that more often from others. I also worry that the difficulties remaining after technical barriers fall will be blamed entirely on a person's motivation.

How Do We Address Differences After Distributing the Tools?

At first, I said people's temperaments make a large difference to their actions. AI responded that temperament is not the only factor.

Do they have time and an environment in which to experiment? Does anyone see what they make? Can they try again after failing? Its view was that these conditions may also shape what happens next.

For example, someone using the first thing you make might encourage you to make another. Conversely, even if you are interested, it can be hard to continue without time to experiment or someone to consult when you get stuck.

I think this point matters if we are to avoid explaining differences in action solely through personality.

AI also proposed education that gives people experience choosing a problem, building something, having others use it, and revising it through to completion. Beyond learning how to ask AI questions, this might provide experiences that lead to another step.

But we cannot yet say that this would close the gap. Requiring everyone to create something or earn money could turn a setting meant to support action into another burden. There may also be cases where AI helps with starting itself, enabling someone who previously could not act to begin.

So we cannot easily decide which way differences across society will move. Still, the question remains: is simply making tools accessible enough?

After receiving the same AI, one person keeps playing and another stops. If that difference extends to whether they have systems that support their next action, equalizing the starting point does not equalize everything that follows.

As the barrier of not knowing the technology falls, could being unable to act become more burdensome?

That frightens me because I, too, have times when I cannot act.

The Exchange That Started This Article

The following is a readable summary of a conversation on September 29, 2026. I selected the parts relevant to this article and expressed the image-generation example in general terms. Predictions and proposals in AI's responses are distinguished from verified facts about society as a whole.

Me:

I think the gap in taking action could become enormous. I'm the same—when I can't get myself to act, I just can't. Even when I understand the reasoning, I sometimes think, “Oh, okay,” and stop there.

Trying AI myself gave me a frightening sense of how serious this was. At the same time, I thought of it as a toy box.

Give people the same game, and some will play deeply and master it while others play a little and stop. Even with the same tools, I think temperament makes a big difference to action. With AI too, there will be a difference between someone satisfied with making images and someone building a system that keeps generating them, selects good ones, and takes them through to sales. The question is how society responds.

AI:

AI could widen the divergence that occurs after people receive the same tools. Being able to consult it where a lack of know-how previously stopped you may make it easier to continue trying, revising, and considering what to do next.

However, I think environment and initial experiences of success matter alongside temperament. Having someone use what you make and an environment where you can continue after failure may also affect subsequent action.

One possible social response is education that goes beyond teaching AI use to let people experience choosing a problem, making something, having others use it, and revising it through to completion.

About the Use of AI

The ideas and experiences are based on the author's conversations. AI was used to check the original conversation and organize the structure, prose, and dialogue summary. During article preparation, I expanded the explanation of accumulation through automation based on that conversation and added related research after checking its publicly available abstract. Claude also reviewed the article's logic and wording, and necessary corrections were incorporated.