The Future of AI
The wrong people are in charge of AI.
That might be an incendiary comment among those in the tech sphere who regularly trumpet the incredible advances in AI we are seeing every day, but I think deep down they know it’s true. I think even the actual people in charge of AI might secretly know it’s true.
What is the purpose of AI really? On its face, the goal is to allow machines to develop the capabilities to learn to think in a human-like way so they can ultimately perform human tasks. Sounds great. What human wouldn’t love a few tasks taken off their plate? Tasks are the worst!
Here’s the thing though, the people in charge of AI have it focused on all the wrong tasks. The “tasks” they are so actively focused on revolutionizing are all things that we do to avoid doing actual tasks. Spend a few moments thinking about how much of your time you spend everyday avoiding tasks. Did you just avoid thinking about how much you avoid tasks because it felt like a task so you checked your phone instead? Exactly, my point.
Let’s talk about one of the biggest “problems” AI is trying to solve, specifically improving internet search through answer engines like ChatGPT.
The powers that be in the world of AI want to revolutionize search so that you can get to the answers you’re looking for faster while clicking as few links as possible. That seems admirable, but when you think about your search engine usage, do you really want it to be fast or would you like it to take a little bit longer. Not minutes or hours, but when you’re searching for a new recipe or the name of the Bonus Jonas Brother, are you doing it because you need that information right this very second or do you need it because you are putting off writing a lengthy email to explain to your boss’s boss’s boss why the numbers from Q2 are skewed due to a technical glitch in your programmatic advertising software’s API connection to your reporting platform?
Let’s be very clear: No one was having a problem with search. Whatever studies these tech companies might be citing are dubious at best. No, the real problem that people are having is that their boss’s boss’s boss’s boss is in charge of AI and thinks we should all be getting back to the very unpleasant task at hand as quickly as possible.
Now I’m sure you’re thinking that generative AI could write this email that I’m putting off which would give me more time to find out if any Major League Baseball player has ever actually batted 1.000. Well, first let me take a minute to thank you for the segue. I was just about to get into generative AI and wasn’t sure how I could change topics and you came through in the clutch. I really appreciate it. Bravo!
Yes, I could put all the salient facts into a generative AI platform that would then turn my thoughts into paragraphs that I could edit and pass off as my own email, but it’s really the editing process that’s the nightmare. Finding the exact right tone in a business email to a higher up is a minefield. Let’s say you have to report some bad news. First, you have to accept responsibility for whatever happened while still making it very clear that it’s not your fault. Then you have to assign blame to some unforeseen circumstance, but make it clear that you not being able to foresee said circumstance was not a personal failing on your part. Then you have to explain how you will put a process in place to ensure that the situation will never happen again while defending the current processes that are in place, which are working the way they are supposed to and that this was just an outlier. And you have to do all of this while making sure you make the person you're emailing feel very smart about something they know nothing about. I don’t care how much time the various AI platforms are logging to read every little thing I’ve ever written, it is just never going to get to that level of nuance.
That’s because the People Running AI don’t have to write emails like that. They have to write the sort of emails that generative AI are programmed to write: blunt, technical emails that try to be vaguely positive, but ring false to the person actually reading it. That’s because these are tech people who became executives without going through the crucial step of spending most of their 20’s and 30’s figuring out how delicate upper management can be and how much energy must be spent to try to keep them placated.
On top of this, once you get a taste for people sending delicately worded emails to you, you’re not going to want to give that up. Think of the sheer power you would feel knowing that your underlings had to spend hours crafting an email to you. It’s got to be an intoxicating feeling, so you’re not going to want to program an AI to do that for people. You’re going to want that to be the only thing your employees are doing to avoid other tasks. The People Running AI want your downtime to be spent worrying about whether the email you’re writing is going to get you fired and not horsing about on the internet trying to determine whether or not Viola Davis would be considered a character actress.
Look, I might not be the first to sing the praises of AI, but I’ll at least show, maybe even place depending on the field. There are amazing advances in AI that will help us solve previously unsolvable problems and make the world a better place. The People Running AI have spent a lot of time understanding those problems and how AI can help. What they haven’t spent a lot of time doing is being regular people who just don’t want to empty the dishwasher again. They just don’t understand what it’s like to be a regular joe who has stupid duties at work and chores at home and only gets a fleeting moment’s peace when they are trying to find a GIF of a video that they kind of remember seeing on the local news as a kid.
So, what do we do if the People Running AI, can’t relate to the rest of humanity anymore? Well, I assume now that I’ve pointed out the problem, they’ll come up with an AI solution for it and I’m really hoping I get a piece of the action for suggesting the idea.