AI Hype?

This week there is nothing really new – no new AI discovery or finding, no new Excel tricks or no new cool projects. The only thing happening is politics and I’m staying away as much as I can although there is a lot going on there.

So instead, I’ll touch on some random AI news that I’m getting and the confusion as to whether we are really heading into some kind of job crisis or not.

A lot of top well-known AI researchers seem to feel that AI will wipe out a lot of white-collar jobs (blue collar will come later with robotics) but studies seem to indicate not much AI impact on jobs.

Corporate chieftains say “full torpedo ahead” with AI replacing employees while articles indicate that behind the scenes companies are rehiring employees.

And in the midst of this, employees are burning out due to the need to verifying everything AI dishes out. The hallucinations are real and the AI workslop is everywhere.

And we are in the early stages of this AI revolution. The hype (?) is that the world of work will be vastly different – possibly in a few years. “Get on the train or get left behind!”

AI slop

Let’s start with the AI slop.

In my case, it’s more like AI fakery.

I encountered two videos that had fake well known academics discussing some topic. One was the astrophysicist Neil De Grasse and the other was Nobel Prize economist Paul Krugman.

Now I don’t follow Neil De Grasse, so I don’t really know how he is physically. I know what he looks like, but I don’t know anything about his speech patterns, his mannerisms, the way he moves. And yet, while watching the video I wondered if he was real. In looking at the YouTube biography, I found a little tiny sentence at the bottom about how the content was AI generated.

I don’t know what clued me in but somehow I figured that “person” in the video was fake.

As for Paul Krugman, I had just seen him talk in some podcast a day or so before, so I had some clues. When the video by Krugman Insights showed up in my feed, I first thought, “He’s now doing videos? That seem unlike him.” Just him doing the video was unbelievable.

Then while watching the video, he was talking kind of fast and never paused. He just went on and on and on without ever moving his sight from the camera. He looked straight at the camera the whole time, not once pausing or hemming and hawing.

The day or so before, he would hesitate, he would pause and look off to the side. He seemed uncomfortable. He was funny and interjected jokes, but he didn’t seem comfortable being in front of the camera. The mannerisms were different.

Unfortunately, when I looked at the biography, there was no indication that the “Paul Krugman” was AI generated but I did note that the Krugman Insights had been created just the day before and there were already maybe 8 or 9 videos created in a day.

Now there is a website Krugman Insights, but I don’t think the videos are real.

It’s really sketchy when people use famous people’s images in their videos to gain credibility for their videos and site.

Burnout

I saw a few headlines about employees burning out due to AI. One aspect of it is the increased workload in verifying the output of the AI. It’s 3 years later and we still haven’t figured out how to eliminate hallucinations.

A friend of mine reported that she read an article about software developers claiming that they are now 17% more inefficient with AI due to the debugging required.

Also, I suspect management let go of people thinking that with AI, they can double up the work on the remaining employees. But instead, employees are now faced with more debugging.

Rehiring

A few more articles have cropped up with the story line that companies that have replaced employees with AI are now finding that there are some things that AI cannot do.

AIs are not ready. There are some nuances that they fail to get. And the hallucinations may be getting in the way. Those companies jumped the gun too fast before the AIs were ready.

That is actually good news, but I wonder how long it will last.

Top AI researchers

A lot of the top researchers keep saying that a lot of jobs will be wiped out.

I thought about this and here’s my take on what they are saying: they may very well be right but they are looking at the end stage and not how this revolution progresses. They can see how the AI progresses to the end stage of replacing most jobs. I can too.

But how quickly will that happen?

And what about the pushbacks this revolution may induce?

The revolution may happen in fits and starts thus allowing us time to acclimate to the new world of work and adapt to new roles. I hope that is the case rather than the fearsome story of mass unemployment.

Because if we have mass unemployment, we may be in a whole different ballgame with different human behavior that we are not predicting. It’s one thing to have 7, 8 or 10% unemployment (which is terrible enough and approaches the levels of the Great Recession) but it is another thing to have a seething mass of unemployed people with nowhere to go.

Societal dynamics totally changes.

I don’t really want to think what that means.

We do need to prepare for it because under such conditions of no jobs to be had, our values have to change. Instead of working hard to make money, the value system has to be something else, because there are no jobs. The whole economic situation breaks. We have to start thinking of something different.

Okay, enough random thoughts here.