Garbage is the Future of Work Imitative AI Promises

The Chicago Sun-Times got caught adding an insert that was largely AI produced and filled with incorrect information. The one that caught the eyes of social media was a summer reading list that contained books that did not exist, or associated authors with books they did not write, but there were other AI generated mistakes in the package overall. The Sun-Times did not produce the package — it was a generic insert meant for use all over the country — and they generally handled this well, apologizing and taking responsibility for allowing it to happen. However, this feels like the future of work in an imitative AI world.

This happened because of Silicon Valley. Not just because of the imitative AI aspect of it, but because the Sun-Times makes much less money than it previously did, largely because Google and Facebook now dominate the advertising market and retain most of its profits for themselves. The reason that the Sun-Times, a large paper in a large city, is buying a generic summer insert rather than producing one specific to Chicago is that they do not have the staff for such an undertaking. And they do not have the staff for such an undertaking in large part because tech companies have eaten their advertising market.

As for the company that produced the insert: newspapers are not going to have a ton of money to spend on it. It is easy for that firm to look at the promises of imitative AI hypers and decide that they can use that to stay price competitive. That means they will hire fewer employees to create these inserts. In this case, I believe it was one person. That one person, however, will have a choice: spend the time required to vet the output of the imitative AI system or not work themselves to death. There simply isn’t going to be enough time, even if there were proper incentives, to vet everything that an imitative AI does given the likely time and employee constraints. On top of that, since the goal is to save money, these firms are likely going to pay less under the justification that imitative AI is doing most of the work, since you won’t tend to get people who are experts and/or have the incentive to do their best work.

This kind of poor work is going to be endemic in the coming couple of years, until this bubble bursts or the fad passes. In the face of tech leadership almost unanimously saying that imitative AI is the future of work, and with these firms offering services at significantly below break-even points, there are going to be many, many, many executives who do not understand either how imitative AI works, how their own products work, or both, who will force imitative AI on their employees. Some will do it because executives as a class do not understand that employees are not a cost, they are a revenue center. Some will do it because Wall Street always rewards layoffs. And some will do it because, like the Sun-Times, they are being squeezed by tech monopolies and near monopolies. But they will do it, and we will all suffer for it.

Our economy is going to be filled with garbage. Some companies will have both the wisdom and resources to avoid this trap. Human produced products might actually even fetch a premium. But in most markets, there is little incentive to be the premium brand. In those other firms, then, we will have fewer people doing more difficult work (checking the output of imitative AI requires expertise in most areas, serious leg work in all of them. And if you don’t know how, for example, to find “hot” summer books, then when you find errors, do you keep running the same queries over and over against until you get a good enough list? How does that save time and effort?), with results that will tend toward the Sun-Times insert rather than quality. And, of course, the lack of quality will be blamed in many instances on the worker, not on imitative AI. A machine that the highest level of the company brought in cannot be wrong.

I sincerely believe that imitative AI, as a business concern, is a short-term issue. I think the combination of terrible quality (which is getting worse) and no realistic path to profits means that sooner or later the will fold as going business concerns. But that can take a while — the market, and the hype, can stay crazy longer than you can stay solvent, after all. Smart companies will avoid it, but we know that most companies aren’t smart. Which means we are going to find ourselves living in a world where everything is slightly to significantly worse and where employees are going to spend a significant portion of their job chasing the bullshit of imitative AI and being blamed for that bullshit ruining a product, experience, or customer interaction.

All because Sam Altman and the leaders of Google, Microsoft, and Facebook cannot figure out something useful to do.