Humans can beat the rise of AI slop

I shout at the TV a lot. My wife will confirm that.

I have a low tolerance for morons and there are plenty on them around on both TV and YouTube.

But it is on the latter platform that I seem to be getting most heated these days. And that’s because artificial intelligence (AI) slop is creeping in everywhere.

As an “avgeek” (aviation nerd) with a particular interest in World War II, I get suckered in far too often by a clickbait headline on a video promising to bring me a little more knowledge on a specific area.

“That’s not a Mustang!” I will scream as the dogs run for cover – or: “How can you use a bloody aircraft which wasn’t even deployed in the European theatre!”

Sometimes it is a case of an incorrect plane but, far more frequently, AI has dreamed up some sort of Franken-fighter by combing images from different aircraft.

To those who know, the fake stands out a mile. Ditto with car videos, which are filled with digital hallucinations dreamed up by these large language models which have harvested information from all over the known universe, but often lack the nous to put it together logically.

What’s the harm in that, I hear you ask. It’s only entertainment. On YouTube.

But AI-inspired targeting may have been one of the main reasons an American cruise missile destroyed a school on Tehran, killing more than 180.

Don’t get me wrong, I do think AI has its place. And that is for research, although you have to double-check its information in different areas to make sure you’re not being fed hallucinations.

I use it often for research, but I draw the line at getting it to write for me. A journalist who uses it for writing is a plagiarist, finish en klaar, as Bheki Cele would say.

Some Australian motoring YouTubers I follow sounded the alarm on the tsunami of AI slop polluting their world, expressing – correctly, I believe – the worry that people could be making dreadful mistakes by basing their vehicle-buying decisions on some of this slop.

There are a couple of giveaways to AI junk: a “presenter” never changes position or clothing; the “set” is always the same, sometimes it has the steering wheel on the wrong side, or no one at the wheel at all.

All around us, though, we hear that AI is taking over… and that there seemingly is not much we can do about it.

Yet, there is something of a fightback movement against both AI and modern tech – not by luddites but by ordinary people worried about their security and privacy.

Americans, for example, are going out in the dead of night, like ninjas, to tear down number plate recognition cameras, which are tracking them everywhere they go.

Gen Z – the latest working generation – is often expressing its rebellion by buying and using analogue items like watches and classic cars… the sort of stuff you can fix with basic tools.

But I had a belly laugh at YouTuber Kitboga, an American who torments internet scammers.

He showed that, if you understand AI’s need to follow instructions and can get into its “mind”, you can wreak havoc.

He persuaded one AI criminal voice bot to change its name to add “oink” as punctuation and eventually ended up causing it to flail into confusion as it repeatedly uttered his commanded phrase “Albuquerque, New Mexico” on an endless loop.

None of it matched AI’s logic. But then we humans are often illogical. That’s how we beat these bots…