Why AI-Made Work Is Often Prejudged as Unfair and Unprofessional
One reason people feel uneasy about AI is that it seems to violate a value especially important to younger generations: fairness. When AI produces something of value, it can suddenly undermine the assumption we have long relied on—that something valuable must have required a corresponding amount of effort, time, or money to create. And when the subject touches directly on people’s own interests, or on something they can easily identify with, that discomfort becomes especially difficult to hide.
There is a romantic idea that as AI becomes commonplace, we will enter an era in which everyone—individuals and organizations alike—simply builds whatever apps they need for themselves. This is often accompanied by the prediction that outsourcing and specialized services will consequently decline. At first glance, this sounds plausible. And for simple products, it often is. A Tetris clone or a to-do list really is easy to build.
The problem, however, is usually not the AI. It is the concept and consistency on the part of the people doing the building. For standardized work with an established workflow, the detailed requirements, guidelines, and accumulated experience already exist. Even when the work itself is highly specialized, that can actually make the problem easier. The opposite is true for individuals with a “romantic” idea of what they want to make, or for groups with limited expertise. They have a much harder time getting there. Their concept and sense of consistency often develop only as the product itself takes shape. As their experience, know-how, and perspective expand, it is perfectly normal for the project to move in a direction quite different from what they originally imagined.
Concept and consistency provide a standard against which decisions can be made. And that standard becomes the catalyst for what AI can accomplish. Give someone a clear formula, and even without AI, they can often produce an overwhelmingly good and polished result in remarkably little time. But the random collection of vague impressions in our heads—the “I’ll know it when I see it” feeling—is difficult to explain convincingly to another person. Often, we are not even confident that we can articulate it clearly to ourselves. How much harder, then, is it for an AI? The AI has little choice but to keep predicting from the short clues we give it at each moment. And, to make matters worse, it periodically suffers from a certain degree of amnesia. That is one reason there is so often a gap between what we expect and what AI actually produces.
This is why outsourcing that can identify, refine, and concretize a concept while maintaining consistency will remain useful. As long as building apps and services requires us to sit at a keyboard and repeatedly persuade an AI, detail by detail, of exactly what we mean, people will eventually become exhausted by the process. Tetris is easy to make because everyone—including AI—already shares a broadly established understanding of what Tetris is. And now virtually anyone can build it. But creating a modern game tailored precisely to your own tastes is another matter. Even with an AI beside you, making it remains a tiring and distant undertaking.
When people are making something, they proudly say that they used AI. When they are selling it, they often prefer not to mention that fact. And when they become the buyer, the moment they encounter the story that something was “made with AI,” they are already prepared to devalue it—often before even examining what the product actually is.
People are accumulating experience in doing something much easier than carefully judging the substance of a product: They see the label “made by AI,” understand that story immediately, and use it as a shortcut to a negative judgment.
