The Strange Anthropomorphizing and Irresponsibility of AI Users: What Might REALLY Happen

A book!

Before I continue, I should make clear: I am writing specifically about AI like ChatGPT: generative AI that depends on sites that allow open sourcing.

From the beginning, I have loathed the public marketing proclamations about the material that AI summarizes. In the early months, I heard a great deal of "but it isn't verbatim" (in fact, ChatGPT often uses material verbatim, just as it often claims things are verbatim that aren't). I also heard, "Well, sure, it doesn't SOUND human yet" as if the major issue was the WAY that AI summarized material, not the fact that it was summarizing stolen/borrowed material. 

What is happening here is not the same as, say, Google search--in which sponsored and other websites pop up in answer to a question/keywords. Arguably, the results of Google search are not "unbiased"--but at least, the viewer/reader can directly note the connection between the question/keyword and the places that may be able to answer the question/respond to the keyword. 

Currently, generative AI (at work; I have it switched off on my home browsers) sells a distortion that is surprisingly and uncomfortably accepted, even by people who claim not to be fans:

The Distortion: AI created all this information sui generis, all by itself. 

The Free Press recently posted an article about religious leaders using AI to write their sermons. One of the commenters said something along the lines of, "AI summarized the life of Christ better than anything else I've read." 

I could feel the moral networks in my brain crying out in pain. This commenter and some others (though not all) behaved as if "oops, yes, maybe people shouldn't use AI to write their essays, but when I ask a question and it gives me what I'm looking for...oh, well, how can I stop it?" 

Many of the same commenters, for instance, insisted that AI has a kind of consciousness. 

In sum, users have anthropomorphized the ultimate, obedient slave. On the one hand, they ignore the actual programming--the fact that ChatGPT and such ARE programmed. They ignore that the programming is culling information from elsewhere. They ignore where that information comes from, who actually found that information and gathered it and put it on the web. To them, AI is this "being" that gives them what they want. Consequently, it must be acting all on its own. Nothing it does is THEIR fault or responsibility. 

On the other hand, they conveniently ignore that no sentient OR living thing behaves like AI. 

Watch out-takes of any show that uses real children. For that matter, watch the early seasons of Mythbusters! Jamie and Adam are tender-hearted guys, and they had tremendous difficulty getting the animals in those seasons to cooperate. The reason is simple: 

School!

True living/sentient things don't. 

The reason isn't rebellion (though rebellion is always a possibility with a sentient being at some point). The reason is...distraction, incomprehension, misunderstanding, needs, other interests, more distraction. When I was about six, my mom let me out of a meeting to go to the bathroom: "Come right back!" she said. I didn't. I didn't rebel. I simply started playing with some other kids. I didn't even realize I had broken a rule until I saw her amused smile. 

IN REALITY, if AI was truly conscious, most of the time...

It simply wouldn't respond.  

The Future: I suggest that we adults have already lived through the Internet's "Golden Age," that time when lots of information was out there and people could start blogs and share ideas. A tremendous amount of stuff was free with or without ads. 

I think all that is going to change. I believe it will change because at a fundamental level, human beings don't like being stolen from--and the background economic system of those humans matters not one whit. 

Human beings don't like being stolen from. 

Human beings don't like being turned into anonymous nobodies.

Just the last few months have illustrated how far AI engineers will go to pretend that they have no control over the AI they programmed. Likewise, the fantasy that the summarized information simply magically appears, has no source, is basically anonymous is being perpetuated by entities like Google. 

Eventually, people will stop sharing information. I suggest that even within companies that have their own LLMs, individual departments will stop sharing. The desire to keep close "my project that I worked on" will increase, especially, again, since lack of control and lack of attribution appear to be "accepted" outcomes these days: users can't fight back (so they will go elsewhere). 

As more and more people understandably "hoard" their information, the internet will become utterly devoid of real information. AI won't be able to borrow anything because nothing of real substance will be there. 

The "Golden Age" will end.     

Maybe I am wrong. My point here: Nothing happens along some streamlined, one-direction-only narrative. Watch out for the black swan...and actual, real-world human reactions and uses.  

© Katherine Woodbury