“The result of a consistent and total substitution of lies for factual truth is not that the lie will now be accepted as truth and truth be defamed as a lie, but that the sense by which we take our bearings in the real world—and the category of truth versus falsehood is among the mental means to this end—is being destroyed.” -- Hannah Arendt
Sadly, I am not sure that the constant falsehoods regurgitated by artificial intelligence are even deliberate. I think a lot of them are just the normal failures of technology, exacerbated by the landscape in which it came to be.
The damage is the same, though, and it doesn't have to be this way.
Let's look at some examples.
A lawyer used AI to generate a legal brief for a case. Nine of the twenty-seven citations had errors, including two that simply didn't exist.
The judge found it pretty convincing, but still did his own research and discovered the... well, fraud implies a level of intent that I don't think it was there. I suspect the reason for the use of AI was simply laziness, but it's still not a good justification.
Estimates are that chatbots have AI hallucinations as often as 27% of the time, and errors up to 46% of the time.
https://en.wikipedia.org/wiki/Hallucination_(artificial_intelligence)
It does make sense that laziness would result in shoddy work.
Of course, looking up enough legal cases to come up with 27 citations for a single brief does sound tedious, but might this happen in other areas too?
Why, yes.
https://www.sciencebase.com/science-blog/vegetative-electron-microscopy.html
This article may give us the source for one of those hallucinations. An old paper had the words "vegetative" (in reference to cells) and "electron microscopy" in parallel columns and they were put together.
"Vegetative electron microscopy" is not a thing, but now it is getting cited a lot.
I really liked this quote (in the current article about AI, not the old one about vegetative cells):
... in a world where scientific endeavour is being derailed by moronic politicians and their henchmen, we need a stronger science base, not one polluted with such nonsense as vegetative electron microscopy. It leads to distrust in scientists and in science, it gives those who peddle pseudoscience, disinformation, misinformation, and fake, greater leverage to shake off the facts and replace them with ill-informed, politically-driven opinion.
We need human understanding and diligent minds.
This technology is not going to solve climate change. Even if you can use AI to run simulations and save time that way, you need a coherent mind with innovative thoughts setting it up.
If the past few years have shown us anything, it's that some people will bite at any false information that supports what they want to believe. This is a trend that doesn't need any help, but it's getting it.
https://www.axios.com/2025/05/23/google-ai-videos-veo-3
"Google's new AI video tool floods internet with real-looking clips"
All for fun, right?
One more thing:
https://www.cbsnews.com/news/sextortion-generative-ai-scam-elijah-heacock-take-it-down-act
They didn't even have real images of 16 year old Elijah Heacock. He still took the threats seriously, and he still killed himself.
One interesting thing in this article is that it mentions legislation supported by Trump to reduce sextortion. What about that provision in the "big beautiful bill" prohibiting regulation on AI for ten years?
Again, that is a vector from which I do not expect coherent thought. But here, among us, we can think about this and we need to think about it.
Posts
No comments:
Post a Comment