Do artificial intelligences tend to swrem left?

[Anonymous]

I asked the question: When creating a story with black characters, is it obvious that right-wing whites will attack?

The answer I got:

"Unfortunately, yes. It is common for stories featuring black characters or that address issues related to diversity and inclusion to be the target of criticism and attacks from conservative or white supremacist groups.

These attacks can be motivated by several factors, including:

1. *Resistance to change*: Some groups may feel threatened by the representation of black characters or the approach to issues related to diversity and inclusion.

2. *Prejudice and racism*: Unfortunately, there are still people who hold prejudices and racist attitudes, and who may attack stories that challenge these views."

Updates

1 y

Do artificial intelligences tend to be programmed by leftists?

[sawno]

Yes, it happens for a few reasons.

Unintentional:
- They want to be factual so they weigh sources like wikipedia higher in the dataset, that control is under leftist control so they automatically learn a left wing view.
- Search engines may steer scrapers to that content
- Useful other sources like Reddit lean left

Intentonal:
- Alignment datasets are used by the big players intentionally givinig them that bias. As hobbyists we do uncensor tunes or abliteration but it can only help so much. Unbiased models have a much higher chance of leaning right though.

Anyone saying they learn from usage is wrong, thats not how the vast majority of LLM's work.

[Mukkiki]

Or maybe it's just reality. Are we going to pretend that right-wingers don't attack any film or animation in which black people are the protagonists?

[sawno]

@Mukkiki No its definitely not reality. Plenty of proof of that. For example wikipedia incorrectly slanders people calling them far right even though in modern times they are moderates and have nothing to do with the far right. Editing is then being blockef so nobody can fix it. Their excuse is that you need a source that they aren't but a video of them explaining their views and denouncing the far right is not accepted. The AI will then copy the far right claim.

Not convinced? Google had their AI aligned so that the image generatiin would be diverse but they screwed up. Someone asked it to generate a picture of a german world war 2 soldier and got a black girl.

Right wingers attacking movies with a black protagonist isn't even remotely the same topic to me. They do have genuine critisism a lot of the time aince hollywood is forcing known characters to have dark skin. Why is snow white a latina rather than a girl with extremely pale skin?

[Asker]

And since when is Latino a race? Latinos can be black, white, Asian. There are literally Latin countries with white, black and Asian populations. And since when does the actress from Snow White have dark skin?

[Asker]

I just disagree with you about Latinos.

[sawno]

she's not pale thats the point. If you are Snow White you want to have an actress white as snow. So a very pale girl. there's definitely going to be actresses that fit that especially if they avoid the sun for a while for the role.

As for the latina part ask the articlea calling her that.

[Mukkiki]

Are you a right-winger? If so, you will obviously deny that it is a pattern that black stories are constantly attacked by right-wingers. No, they are not only attacking stories that involve race swaps, they are attacking any story that has black characters as protagonists, including stories in which the characters are originally black. I can name dozens of books, films, animations, etc., that are constantly being attacked by right-wingers.

Right-wingers still deny that it is a pattern, but as things are happening in a few months it will be impossible to deny.

[jshm2]

Depends on their datasets. Most data is acquired from White and males - as that is what most data is polled, and based around. There have been shifts these past few years to switch that to more relevant sources.

But when using any AI model it's important to always add the "show sources" prompt. Also, a lot of Chatgpt data has been frozen not only for copyright but because it uses old data, and even makes shit up.

[Mukkiki]

Or is the AI simply based on the pattern of right-wing social media behavior? It's standard practice for right-wingers to attack stories involving minorities.

[Mukkiki]

It's just reality. Go into any game, movie or animation trailer and I guarantee you'll find right-wingers making racist comments. It's not the AI that's being manipulated, the AI is just showing what it sees happening on social media.

[strateguy632]

As if biden never called schools with blacks " jungles ". Isolated whites is racism.

[Mukkiki]

@strateguy632 So what? Biden can barely open his mouth and spell sentences. , right-wingers attacking films, series and animations in which black people appear are thousands. I don't even know what Biden has to do with my answer. The fact remains: AIs are not left-wing, they just see patterns of behavior between ideological groups. A right-wing pattern is to attack black or LGBT actors and any movie or show with black or LGBT protagonists. Biden is just one man

[Agape93]

No. They learn and adapt based on the information you feed them.

[msc545]

No. The AI's creation reflects reality. The left wing does not attack blacks - the right wing does. That is reality.