Skip to content

A human-centred process is more important than an AI-tool-centred product

Using AI-powered tools to produce something seems to support a culture that says the end result is more important than the process. Sometimes that’s true, of course. But for many things, I think it’s not. More importantly, it seems to be leading us away from a human-centred approach to things.

The focus of my concern is more on the use of AI-powered tools that on AI itself. How are we using these tools? How much are we weighing up the pros and cons of these (and other tools) before using them? Are we really thinking this through?

Coding

When writing code, the process of figuring things out is what helps the human learn. We don’t learn by just seeing the end result. We don’t learn the material for an exam by just reading all the answers of old exam papers.

I mean “don’t” here in a slightly complicated way. I mean don’t as in it’s not the best way. We don’t learn as much, as well. We don’t remember as much, we can’t recall as much. Broadly speaking, because reading the solution to a problem is passive and figuring it out is active.

Creative work

I think the importance of being human-centred is even more clear for creative work. As an artist, a writer, a creative, the process of doing the work is deeply important. The process of making something changes the work as it’s made, and it changes the one doing it. And making something can be fun!

It’s also important to me as someone interacting with a work. Who made this, why, when. Having some background enriches the experience of the work.

  • “It doesn’t matter who made it” is easily refuted. Imagine you receive a book as a gift. It matters, it feels very different, if it’s from a stranger or from your best friend.
  • “It doesn’t matter how it was made” is similarly easily refuted. Imagine you receive a cake as a gift. It matters if it was bought from a shop or hand-made by a person.

This is not to say that one is “better” than the other. But there is a substantive difference to our experience.

It’s also not to say that we humans can’t use AI-powered tools to make better, more interesting, weirder, cooler, art. But the AI-powered tool should be enhancing and improving the work of the human artist, not replacing them. A person being creative with AI rather than AI being creative instead of a person.

More consequences

Some other things tumble out of the assumption that the product is more important than the process. One is the assumption that “more is better”. A common argument is that because AI lets us make stuff faster, we can make more stuff. It doesn’t feel like “not enough stuff in the world” is a problem we have.

This feels especially true for creative work. Do we want higher quantity? Or higher quality? These aren’t mutually exclusive, but it does feel like current, mainstream, AI-powered tool usage is tending more towards higher quantity as a design goal.

Another assumption that tumbles out is speed. The implicit assumption that faster is better. This can be true when the desired finished state is very clear, and simple. But faster is not often better. Anything that’s meant to be an enjoyable experience is a good example. Listening to, or playing, music isn’t made better by by speeding it up. A good meal isn’t made better by eating it faster.

What do about it

This doesn’t mean “don’t use AI.” It means maybe we should be more careful, be more human-centred about this. Maybe a human-centred process is more important than an AI-centred product.

I’m a techy person, in a techy job. But that doesn’t mean any new technology is automatically something I pick up and use. I keep coming back to something I read in Cal Newport’s Deep Work, about taking a critical stance to adopting new things. He says something like:

Take a craftsman’s approach to tool selection: pick up a new tool if the positive impacts on core factors to success and happiness outweigh the negative impacts.

Take a moment, pause, before enabling or adding AI to everything and anything. Will this actually be helpful? Or is something a bit slower, a bit more human, a better choice?

(Psst! The answer is usually the unsatisfying, but true, “it depends.”)

Things that prompted this pondering