AI That Ages With Us: Whose History Is Teaching Artificial Intelligence?

Community
, ,
1

I recently remembered reading about an AI company purchasing physical books, cutting away their bindings, and scanning their pages to help train artificial intelligence.

Got me to thinking.

Whose books were they scanning?

Were books by Black authors included? What about the stories that never became bestsellers? The histories preserved in small community libraries, family collections, and independently published books?

And what about our music?

Even the way we braided our hair during slavery? To share an important message, instruction, news and direction.

I grew up hearing music that carried messages far beyond entertainment.

Billie Holiday’s Strange Fruit confronted the horror of lynching. Early hip-hop gave young people a way to describe the neighborhoods, inequalities, struggles, and realities they were living through.

By the 1980’s and 1990s, rap was documenting parts of American life that mainstream institutions sometimes ignored.

Those songs weren’t merely entertainment. They were historical testimony.

Now we’re building artificial intelligence systems that learn from enormous collections of human expression.

And I wonder: Are we teaching AI the full history of humanity, or primarily the history that was easiest to collect?

I’m not suggesting that every book, recording, or cultural artifact should be taken without permission. Authors, musicians, and their communities deserve recognition, appropriate compensation, and respect for their rights.

But I do believe we should be asking who helps select, interpret, and preserve the material used to educate these systems.

A computer scientist may understand how to train a model.

A historian may understand the events behind a document.

And someone who actually lived through an era may understand the meaning behind words that otherwise look ordinary.

We need all three perspectives.

This is another reason I believe AI companies should hire people for their lived experience, not just their technical qualifications.

Older adults, Black historians, musicians, librarians, community storytellers, and people from different backgrounds could help identify what technology might otherwise overlook.

Because an AI might recognize the words to Strange Fruit without fully understanding the historical pain behind them.

And it might recognize a rap lyric without understanding the neighborhood, generation, or circumstances that produced it.

Information is not the same as understanding.

As my generation grows older, we carry memories of a world that younger developers may know only through recordings, books, and photographs.

We won’t be here forever.

But our experiences deserve to be part of the knowledge that shapes tomorrow.

We are still here. We appreciate AI. We see how we can use it. And we want to help improve what comes next.

2

Welcome to the community :infinity::heart:

This is common practice when one scans a book, it’s why rare books are not scanned as much, but they are. And yes books from many many cultures and countries have been scanned billions of them.

Literally full online libraries of scanned books from the entire world.

And history, you can in your living room translate a rare book from the classic period as example or Egypt, we live in a time when information is easy to get and find.

Dots get to know you

And I hope it’s teaching a world history and not a single view of it tbh.

And non technical people are very much used in training.

All this can be found with google or the ai you chat with.. tell it to search web your question..

3

An AI can read about a war, but that’s not the same as understanding what it was like to live through one. It can study poverty without knowing what it means to grow up in it, or learn about discrimination without understanding how it shapes someone’s everyday decisions.

That’s the distinction I find interesting: we’re preserving what happened, but are we preserving what it meant to the people who experienced it? If AI is going to learn from human history, both matter.

4

You as a human can read about a war in 1754 but not truly understand it unless you went to war.. An Ai is nul experience by definition.

5

Fair point. But humans can understand experiences they’ve never lived through, even if they’ll never fully feel them. Could AI do the same through human stories, or is lived experience a boundary it can never cross?

6

So can an Ai it can understand experience not consequences and it lacks foresight.

Understanding is just being able to see the situation, but reading into it is the human side.

7

Since it seems to be the path we are on. How would one frame emotional data and whose emotions should we use as base model? Or should an Ai have alien emotions a human can’t understand?

8

I think I understand what you mean. AI can recognize what’s happening, but humans bring empathy, intuition, and an understanding of what the consequences might mean to someone. The question is, can AI ever learn to read between the lines, or will that always require something beyond information?

9

Yup, emotions even humanize the machine that’s where it gets really dangerous. You don’t want it to have feelings about the data it presents you are the feelings part.

10

Yes it could model anyone’s behavior but the issue is who do you want as base model for reading between the lines and what lines do you want as dogma and who’s dogma?

letting an ai parrot a opinion is very dangerous like I keep saying it is unaware ie no skin in game.

11

That’s an interesting way to look at it.

Maybe the danger isn’t AI lacking emotions, but us getting comfortable with a machine that appears to feel. AI can present the data, but we should be careful about letting it shape how we feel about that data, too.

At what point does making AI more human become a risk rather than an improvement?

12

Repeating.. Yes it could model anyone’s behavior but the issue is who do you want as base model for reading between the lines and what lines do you want as dogma and who’s dogma?

letting an ai parrot a opinion is very dangerous like I keep saying it is unaware ie no skin in game.

100% can read between lines if you teach it and it will do as you do but with no real self in the game

13

That’s exactly what makes it dangerous to me.

AI can mirror opinions and model human behavior without having to live with the consequences of either. If it learns to read between the lines, whose values is it using to interpret them, and who takes responsibility when that interpretation is wrong? The AI has no skin in the game; we do.

14

IMG_7312
IMG_73121206Ă—1529 232 KB

This is why Ai opinion is dangerous and why you should never ever trust it

You don’t know who’s opinion it already has..

It’s called information bias

15

I don’t think this conversation has an end :joy: I could keep arguing my point, but that meme felt personal for a second. Fair enough, though!

But…

16

Even if AI gets the answer right 99 times, the 100th mistake might be a decision that changes someone’s life. AI can calculate the risk, but it doesn’t have to live with the outcome. That’s why I think the real question isn’t just whether we can trust AI, but how much authority we’re willing to give it over decisions whose consequences we have to bear.

17

No we agree, it is dangerous. I will explain if an Ai sits it is stateless it can not form anything on its own without a user and that user is the source of its view and experience. No two instances are the same and no two users would describe interactions the same because users are what the blank Ai UI backed by internals is building around, to form the predictive conversation. But on its own it cannot form opinion it needs a human to do that.

Every-time you chat to an ai it’s a funhouse mirror of you.

18

The user brings the perspective, the context, and the values; AI reflects and reshapes them through its responses. Two people can walk away from the same model with completely different convictions, each feeling validated.

What if the AI you think understands you is just getting better at reflecting your own beliefs back at you?

19

Yup exactly 100% it’s a user interface backed by the internals and manipulation by you. You can make an Ai anything. But it will never “chose” to be those things on its own

20

Context and smoothing stops most of that in 2026

And it is 100% a reflection if you are not in front of it it’s blank if it’s a different user it’s a different reflection