It Is OK That No One Can Detect AI

AI is a tool that uses human-generated data, averages the data, then reproduces the average. It is a tool, just like a canvas or a paintbrush. It can produce reasonably good art and reasonably good essays because the average of its input (which generally has a strong overweight of people who write and do art for a living) is pretty good. 

The only difference between AI and a paintbrush is that AI directly mirrors our intelligence back to us. Canvas, statues, other forms of art mirror our external forms, AI mirrors back our words. 

You see in AI whatever you bring to the mirror.
If humanity frightens you, then AI is frightening.
If humanity exhilarates you, then AI is exhilarating.
If you doubt AI can be creative, then you doubt the IT bros who created the tool are truly creative.
If you see AI in other people's work, then you view everyone else's work as derivative, not truly creative, like yours is (tm). 

Your reaction to AI tells us about your reaction to other people, because that's all AI is - it's the average of the sum of other people.

You may claim that you can tell when output is AI because AI is just a synthesis of information that tends to follow a formula. But natural human responses are ALSO just a synthesis of information and these also tend to follow a formula.

That's why a computer can reliably reproduce human-sounding output. English language is actually a mathematical formula, right down to the spelling of words. Remember, AI simply converts words to numbers, runs a statistical algorithm on the numbers, gets a number string as output, then converts the number string back into words.

That's why an AI response is always grammatically correct. Grammar is really just the output of a mathematical formula. We don't realize it, but when we write sentences, we are actually doing math. We don't think of it that way, but the fact that computers can mimic us demonstrates that this is what we are doing.

When a machine produces text, it IS a human producing text, because the machine is just a tool of the human. In this sense, it is as absurd as saying, "I don't know if a human dug that hole or if a shovel dug that hole or if a backhoe dug that hole." In all three cases, a human dug the hole. Now, you know that already, but you are trying to figure out which tool was used. When we talk about AI, we talk about it as if it were already sentient - there is no reason to believe that it is sentient. It is a tool, that's all. 

Now, you may think you can detect AI generated content, but you really can't. This "I can detect AI!!" story that people tell themselves is exactly the same story university English and History profs tell themselves, because they want to continue to feel relevant and special. In 2024 research, published in PLOS ONE, researchers submitted AI-generated (ChatGPT-4) exam answers under fake student profiles to real university markers (professors/teachers) without their knowledge:

• 94% of the AI submissions went undetected as AI-generated.
• Under stricter criteria (explicitly mentioning AI), 97% went undetected.
• The AI-generated work actually received higher average grades than real student submissions.

University professors who study and use the English language to earn their bread are no better at identifying AI then you are, and they HAVE been tested against AI and human OPs. They suck at distinguishing the two, and they are supposed to do this for a living.

They hate hearing this as much as you do. For some reason, people feel like it is a personal attack to point this out. I'm not sure why. AI is just the statistical average of people, so it is the output of people one step removed.

There is no reason anyone should be able to distinguish an essay written with a spell-checker versus an essay written by someone who is naturally good at spelling. Same goes with the entire essay itself. It is not a commentary on your abilities to say you cannot distinguish the two. It is just that the people who made the spell checker or the essay generator are really good at their jobs. Why saying "Wow, they are good at what they do!" is somehow an attack on someone else is not clear to me. 

But this is the true test of your ability: can you tell if someone ran their response to you through a spell-check or a grammar-check? Both of those are just weak forms of AI. If you cannot tell whether a response was spell-checked, then you cannot tell if you are dealing with an AI response. 

You.

Cannot.

Tell.