Ethics as Lightning Rod

I’ve read many books about AI in recent years. Almost every one of them has a section on ethics. Usually toward the end of a chapter. It names a problem, urges caution, and then the next chapter starts as if nothing happened.

I tried to derive something concrete from these sections. An action item, anything that someone would do differently tomorrow. I found nothing.

The phrasing is interchangeable. “It is important to consider ethical aspects.” “Responsibility ultimately lies with people.” You could move the paragraph from the marketing chapter to the HR chapter and it would work just the same.

In electrical engineering, there’s a component that discharges voltage before it can cause damage. A lightning rod. The ethics sections in business books work the same way. They take the tension that could arise if someone asked: Is what we’re doing actually right? And discharge it into a paragraph. After that, you can carry on undisturbed.

This is a clear pattern, and business books need these sections because a book about AI without ethics looks reckless. So it gets inserted where it belongs. Just enough to mark the claim.

When we were writing our own book about AI, I sat in front of the same problem. You’re looking at the manuscript and you know you have to say something about ethics. And then you realize how hard it is to say something that goes beyond a warning. Something with a real consequence. Something that doesn’t let the reader just turn the page. Many authors solve this by not trying in the first place. I would have preferred that too. I’m not happy with my own result. Not at all. Why else did I spend that time writing one essay after another about it. Always with the same question and a similar result.

Anyone who reads these sections walks away feeling the ethical question has been dealt with. You can move on. But the real questions are nowhere to be found. What happens to the people who lose their jobs to AI? Where do they go, who pays for retraining? And what about decisions that nobody can trace anymore because the analysis came from a machine?

These questions can’t be discharged in a paragraph. They would call the entire book into question. And that’s exactly why they’re not in there. And they’re in here because I’m dissatisfied with my own work. Could I have solved it better? Probably not, otherwise I wouldn’t still be wrestling with solutions two years later.