So, back in ye olden days, a classic tactic to manage how many complaints or requests a department received was to just make sure any requests had to come in with a filled-out form. The very act of making someone fill out paperwork was enough to discourage many people from even bothering to try.
Say what you will about LLMs, but one area where they excel in is “doing stuff lots of other people have done before.” They’re also great at parsing through official-ese to identify what steps need to be done to complete a procedure. The net result of this is that departments are seeing a huge spike in the number of requests they’re getting. This is a double edged sword.
On the one hand, it’s made it easier for people who have genuine needs to get those needs met. I used it myself a year or so ago when I was trying to get an airline company to do right by their stated policies. They were giving me the runaround, and AI helped tell me what paperwork I needed to fill out, where to submit it, etc.
On the other hand, it’s got the potential to overwhelm departments that are already underfunded as-is. It’s one thing when it’s a big business that ought to have better customer support to start with. But when it’s your friendly public library, and people have decided to put in a whole stack of book challenges they get from an outside organization with an axe to grind? Add into the equation the ability for agentic AI to merrily sit there filling out paperwork on the behalf of anyone asks, and you’ve got the potential for serious abuse.
What’s the answer to this? As with many things, I don’t know. There’s part of me that really likes the ability for people to cut through all the red tape by using an LLM to get the thing done that they wanted to do. But just because something’s easier to do doesn’t necessarily mean that it should be done on large scale. Part of that red tape seems to have been built on a premise of good faith. Since it was hard to fill out, people only filled it out when there was a real need, and then departments would look at those requests as they came in.
That’s not likely to be possible if this continues. I suppose we might end up in a world where AI is filling the paperwork out and AI is reading through it and dismissing it. If that’s the net result, then we end up more or less where we began. It would be nice to think people would pinky swear to only fill out paperwork they really cared about.
Anyone wanna bet that happens?