Why does the future speak in old words?

Call a machine an assistant and you have already given it a place in the org chart.

We still dial a number, though nothing turns. We hang up a phone that has no receiver. We put files in folders on a desktop that isn’t a desk.

The objects went. The words stayed.

That isn’t stupidity. Familiar words let us approach something unfamiliar. You don’t need to understand how a computer stores information to know roughly what to do with a folder. The future often gets through the door dressed as the past.

But I think something different happens when the old word names a role rather than an object.

An assistant helps with a job you have already defined. A copilot sits beside someone who is still flying the plane. An agent sounds as if you might hand it the controls. Those names don’t just describe software. They suggest who gets to think, who gets to make, and who has authority.

I’ve spent more than 30 years in creative work. I’ve seen agencies organise around expertise that was difficult to access: who could cut the film, build the experience, produce the sound, or turn an idea into something an audience could feel. Those specialists bring judgment, not just technical operation. But the shape of the organisation grew around the scarcity of their time and tools.

That is why assistant is not a neutral word in a creative business. It slips neatly into the existing chain. Someone decides what the work is, then asks the software to help them do it faster. It can draft, tidy, resize, summarise. Useful, certainly. But nothing in that description asks whether the person who first saw the idea can now explore its form, or whether an expert might be more valuable shaping the work than guarding the first step of making it.

Calling the software an agent doesn’t fix this. That word can imply authority the system has not earned. It may carry out a task, but it does not live with the consequences of a weak idea, a misleading image, or work that leaves an audience cold. The ease of making a first version is not the same as knowing which version deserves to exist.

So I’m less interested in finding the perfect new noun for AI than in asking what our nouns let us avoid. A business can buy “assistants” for everyone and leave every creative permission exactly where it was. It can call them “agents” and mistake automated output for a creative decision. Both are ways of adopting the technology without deciding how the work should change.

Here is the answer I think matters: stop naming the machine as if it has a job title. Name what people are newly able—and allowed—to do. Who can explore an idea across disciplines? Who can show a first version? Who brings the taste to know it is wrong, or the craft to make it right? Who remains responsible for what reaches an audience?

That is a much harder conversation than choosing between assistant, copilot and agent. It is also where the creative opportunity sits. The words help us welcome a new technology. They should not be allowed to write the old organisation’s rules for it.