The Future Contradicts the Present

The AI debate has two halves and the people leading it probably see it differently. But if you lay the arguments side by side, they tell two different stories.

The first half is about the partnership between humans and AI. AI supports by taking over tedious routine work and delivers the data that humans then interpret. The story goes: AI delivers, humans decide (mostly, still).

The second half sounds different. Here, AI starts acting more independently. It develops strategies and decides on customer offers. Humans then give “impulses”, they “accompany” and “refine”. You can tell from the choice of words what has happened. In the first half humans act as agents, in the second they only react.

The whole thing is a seamless, invisible drift from “AI assists” to “AI takes over”. So slow that you don’t notice it, even if you were paying attention.

I think most people didn’t notice either, because the transition is barely one, because productivity from the outside simply keeps improving steadily. AI implementations work exactly this creepingly in practice. First a pilot project, then all the small tasks get handed over. Clear boundaries are drawn and it is supervised by humans. The tasks then grow, the boundaries shrink and the supervision is no longer needed in many places, because the trust is there now. At some point the machine does it completely and nobody asks about it anymore, because there are more important things to do than things that have already been outsourced.

This isn’t a new pattern. It’s the same with every technology that has to prevail. It starts with the promise of efficiency gains with full control over the output, whatever that looks like. Because the machine is supposed to be just a tool. The better it does the work, the more the results improve and the whole process becomes cheaper as well, and the question why a human should still do it doesn’t come up anymore. The machine takes over and that’s just how it is. The human is replaced.

What bothers me about the debate isn’t that it describes this development. What bothers me is that it does it so matter-of-factly. At the start there is the assurance that humans stay at the centre, at the end the matter-of-factness that they no longer are. Both from the same people, without them even beginning to name the contradiction.

The reason for this is that the debate wants to be both things at once. It wants to reassure and excite, to say that everything stays the way it is and at the same time that everything becomes different. It wants to address the manager who is afraid of losing his people and the manager who dreams of cutting his costs. And because it wants both, it says both, without noticing that it doesn’t fit together.

Depending on which of the two halves you believed: the beginning is meant to work like a sedative, the end is the new reality as the only logical development that everyone went along with.

I prefer the more honest of the two versions. And it goes like this: you cannot predict it and you do well to say so openly. Strength is describing the range of possible outcomes and their consequences clearly and developing alternatives for people. A discussion that thinks things through to the end helps everyone. Right now, those who expect an honest answer are fobbed off with two stories that don’t fit together and those who notice it are becoming more.