AI Slop vs the Human Connection

After a hard day of work or study, some people enjoy a mindless scroll through social media. It’s an opportunity to turn one’s mind off for a few minutes of harmless dopamine, right?

Oh dear. Given my job and search history, the social media algorithms think I must really want to see some AI slop about physics education! Here’s a few examples I have seen after a quick 5-minute scroll.

What is that?!

This little image was on a page of AI slop notes about Gay-Lussac’s law. If anyone knows what it’s supposed to be showing, please do let me know! Nonsense like this may be harmless, but some of the examples below are worse.

I can only count to four

Any idea what the other particle in the nucleus of the H-2 and H-3 is? I’m no expert in mathematics, but one proton plus two neutrons equals three nucleons, so why does H-3 show four nucleons?

Gay Lussac’s Law leaving us in suspense

If the temperature increases… what? If the temperature decreases… what? Can the pressure not decrease? What is this trying to say? In case you are wondering, there was nothing else to the right of this image. It was just this. Plus, if this is supposed to show how pressure depends on temperature for an ideal gas of constant molar amount and constant volume (and that is what the AI slop poster said it was), then the line should be straight.

The unforgivable crime

Students and colleagues of mine will known that few thinks irk me as much as when Newton’s Third Law is incorrectly stated. The ubiquitousness of confusion about Newton’s First Law vs Newton’s Third Law would be funny if it wasn’t so damaging. The worst possible wording is “every action has an equal and opposite reaction.” This statement uses the technical turn “action” incorrectly (action is the integral of the Lagrangian of a system over time). It also implies cause and effect, where Newton’s Laws are classical physics an d the two forces are simultaneous. But perhaps I could overlook that if it weren’t for the diagram showing the “reaction” force being the “upward lift”; the reaction to the “thrust”. What? Just… what? Action “kick”, reaction “kick back”? I think need to increase my blood pressure medication and have a little lie down.

What’s that blue arrow?

Maybe it isn’t explicitly stated that the blue arrow is the amplitude, but it’s implied, and a student could be forgiven for thinking that’s what the notes show. Also, there isn’t really enough of the waveform shown to clearly demonstrate the wavelength being as indicated, but given what is to the right of the diagram, I am alarmed that anyone thought this would good enough to include in their notes. Ah, but no person did. This is good enough for AI, which leads me on to…

Well that clears that up!

Longitudinal waves and compressions and rarefactions. What else do you want? What, you want to know where they are? What they are? How unreasonable. Just be glad they’re written on the diagram somewhere.

Crest and… wait… what?

Where is the trough label here? Yet another example of a “that’ll do” attitude that AI seems to have when it comes to physics diagrams. There’s evidently no real understanding of the very basics of physics here. And that’s clear when looking at the last image.

The pen bends, what more do you want?

Given how light refracts when passing from water to air, the pencil would look like it’s in shallower water, not deeper. The AI knew light appears to bend, and it bends towards the normal when entering water, and pencils look bent when in a glass of water. It put it all together just as a child might when learning about this for the first time, but without the critical thinking it has got it wrong.

Conclusion

All of these examples were found in the wild (on Facebook). I wasn’t looking for them, Facebook thought I would want to see them because of my interest in physics. In all these cases, an account clearly (hopefully) controlled by an AI bot posted these AI generated notes, and AI operated accounts gave the posts an abundance of “likes” and generic comments like “thanks” and “so helpful” and “love it!” All of which feeds the AI algorithms, and makes them favour the AI bot accounts.

I sincerely hope that humans doomscrolling through social media are clever enough to ignore the AI slop. I think it is a good thing that the perception of AI as being a “useful tool for the future” is falling away, and is being replaced with “actually, it’s a bit rubbish isn’t it?”

I have some big concerns about the proliferation of AI. Maybe it is a caricature that those in charge who hoard the wealth are eager to use any and all means necessary to increase their profit at the expense of the very people who made them rich. Maybe it’s an exaggeration to say that when the majority of the population are put out of work by computerised systems designed to be cheaper to operate than humans, that the humans will revolt. I do think there are bigger issues to be considered when it comes to the rise of AI tools. Real “meaning of life” stuff. I’m not going to get into that in this post. I’m not really qualified, and I’m still recovering from the stress of seeing the AI slop this evening.

And don’t get me started on computer systems “apologising” when things go wrong (can they feel the emotion of remorse?) or “thanking” me (can they feel the emotion of gratitude?)

Humans have the capacity to feel gratitude, and I feel tremendous gratitude to all those who support what I’m trying to do with Physics with Keith: providing high quality physics education for all.