From the Editor’s Desk: Slop ⭐

AI deniers and critics have latched onto the term “AI slop” for obvious reasons, but human error is a far more serious problem, one that has led to countless disasters and untold loss of life. If you want to remove “slop” from the world, then the goal should be to automate everything with technology. And that technology will most certainly utilize AI.

Some of the best (worst?) examples of human error include the sinking of the Titanic, the Chernobyl nuclear disaster, and the Exxon Valdez oil spill. But human error leads to disaster literally every single day in the form of automobile crashes. It is by far the biggest factor in airplane disasters and maritime accidents. And it plays a significant role in medical procedural errors too. How many sporting events have been decided by a missed call from some human referee or umpire?

Too many, and that’s why each sport is increasing the use of video reviews. But let’s keep this to technology.

From the CrowdStrike disaster to almost every single Microsoft security incident we’ve seen, it usually comes down to one thing, human error. And if you factor in indirect human error, as in the coder who wrote the buggy software that failed, as we should, then it’s even worse. Inserting a person into the chain of decision making is arguably one of the biggest mistakes one can make.

And yet, that is what we’re calling for with AI. It’s like the feeblest person in the room demanding to make decisions for everyone else.

We need to let go.

I have been a car enthusiast for my entire life, and I stuck diligently to small, manual-shift vehicles for as long as I could despite the obvious downsides. I still believe that I’m a better driver than virtually everyone I encounter on the roads, especially here in Pennsylvania. But with our streets and highways now crowded with traffic at all hours, it’s pretty clear that no combination of laws, enforcement, and collective driving skill will ever solve this mess. What will solve it, of course, are autonomous vehicles. Which is precisely what an enthusiast like me would resist. Sometimes it’s difficult admitting that you’re part of the problem. We're so human.

To head off the obvious counterpoints, yes, AI still has a way to go. I’m not sure that anyone watching the first Wright brothers flight could or would have foreseen prop airplanes, jets, or spaceships, but they certainly would have gazed in wonder at a thing that they had previously thought to be impossible. Today, you don’t need much of an imagination to see how AI will transform our lives. In some ways, you need more of an imagination to claim that it won’t.

No matter. Despite ongoing improvements, AI will still make mistakes, it’s inevitable. As we rely on it more in ever more critical ways, lives will be lost. And each time this happens, the deniers and critics will explode like cockroaches from under the wood pile and complain bitterly that this is exactly why they oppose this change. Which...