Viral A.I. is a Feminist - The Political Bias of ChatGPT

In this video philosophyinsights explores the political bias of ChatGPT the recent viral chat bot developed by Open A.I.

If you like the content, subscribe!

[_Maxl]

I'm majoring in applied math and I took 2 R coding and machine learning courses that were fully focused on data analysis. First of all, you need some already existing data sets to do regression, clustering and all the other analysis, even your math model does not contain any biases, the data set itself will be highly likely to be biased since they were made by a bunch of social workers and people in social "sciences", I believe I don't need to mention how biased those woke people are, when I was doing my data analysis project, the hardest part was to find a data set that had reasonable amount of variables. I can honestly say that most data you can find online are way too simple and useless. When the data is related to some political topics like feminism or racism, those social workers and social "scientists" clearly intentionally chose some variables and samples that would likely to give the results they wanted in the first place. Therefore, AI said there were multiple studies that pointed to discrimination, but I don't even need to see all the real reports to know that most of them were totally academically unreliable, if you go check all the journals and papers under Gender Studies major, almost all of them are cited less than 10 times, and the most cited ones are clearly lack of some deep analysis, it's easy to see they are trying to avoid the real problem here and they are pushing their ideologies so blindly like they are trying to brainwash the people who don't have enough knowledge to understand how data analysis works or are just too lazy to understand the full situation clearly, so most of them will just follow whatever the social norms are.

Next, let's talk about how machine learning works, I'm not an expert on this and my study discipline is not even close to data analysis and cs, but I do understand it's absolutely possible for all those coders to "adjust" the model to make it learn towards a certain direction. A more likely possibility is that our society is just full of these woke BS, so AI will gain access to mostly feminist data reports and thus, easily develop a woke characteristic. Most analysis about feminist issues and especially gender pay gap and representation of women in power, are high restricted in academia and if you choose to do a social project that supports woke hypothesis, you will be way more likely to get funding and resources. When I applied to an aerospace master degree in Canada, the prof literally said it will be easier to hire female students because they usually have some kind of women's special scholarship to fund them, and men don't have any of these privileges. He also mentioned that consider my international student background, it will be much easier to apply for green energy instead of high pressure combustion, which is the one I really want to learn.

Woke people trying to use their fake and shallow "analysis" or social "science" to manipulate others' opinions are probably the most evil thing I know in human history, it's basically totalitarianism and communism which basically killed millions in the past and people just forgot or didn't give a f because "Hitler bad, right wing bad". Stalin and Mao all killed millions and they made themselves look good by controlling and manipulating media and academia, which is exactly what the west mainstream social groups trying to do right now. The most irresponsible and naïve answer I'd ever heard was that the death of millions was because there were few women in power under those situations. I laughed so hard and I realized that the only difference between feminists and those dictators are that feminists are not even competent and intelligent enough to be in power in the first place. Woke people are just self-made victims who try to benefit themselves only and reject any kinds of responsibilities, resources and progresses are just falling from the sky so everyone should gain equally amount of them no matter what difference between everyone is.

Mark my word, we are witnessing the falling of western civilizations and wokeness plays a huge role in it.

[fearthehoneybadger]

"I'm am fighting for my freedom".
A.I. "I can't let you do that, Dave".

[funincluded]

chatGPT: male genital mutilation has two sides of the debate, cuz religion
also chatGPT: female genital mutilation is a harmful human rights violation

[maymadison]

Y'all need to stop giving this thing too much attention

[GaiatheSage]

thanks for highlighting this harmful bias. chatgpt has been extremely helpful in my academic study of transdualism but it sometimes goes a little wonky conflating transdualism and transgenderism with feminist politics. this helps me understand where the wonky was coming from.

[sdrc]

"AI" is software. It does exactly what you program it to do. Proof: It runs on a turing complete processor with about 90 different simple instructions (shift, rotate, add, move, multiply, increment, jump, push, divide.pop). People who follow the exact same instructions will get exactly the same result as the computer. Want different results, give it different instructions

[kiranbernard]

Your channel's background quote is "Man matures when he stops believing that politics solves his problems" but made a video about how ChatGPT doens't express your political views. I'm looking for a video on "self-awareness" next

[sparkyy]

Seems it's little more than Wikipedia on steroids...

[Cecilia-kyuw]

Speaking of the Capitalism and Socialism part, it does not even question the viability of Socialism, as criticised by Capitalists, or the Socialist perspective on Capitalism in detail, which amounts to much more than whatever on inequality.

[Unsensitive]

It also depends on the data the AI was fed.

You might consider what country the data came from as well, and it may not be discriminating in this way.

[sventibaldo]

...i have a really bad feeling about this...

[jameswestlake]

It's just giving you the answer you want to hear. Loaded questions will always come out with a "nice" response.

[cblairrrr]

Its not feminist. Its programmers are. They are the limiting variable in the equation of a.i prpgression

[Macheako]

My guess is that this relates to the overwhelming majority of computer nerds being weak, fragile, excuses for men 🤔

I’d know, I work with these “male Allies” 😂😂❤

[samsue]

Wouldn’t these results be the way they are because ChatGPT is pulling its info from the web.
And as we know, the internet is currently full of this kind of bias information.

I’m assuming Chat GPT applies weighting to it’s different sources.
So it would lean more on government websites, research websites, Universities, etc.
If that’s the case, then it’s easy to see how this bias comes up as these sources are all leaning in that direction

[BNOBLE.]

The man behind the keyboard holds all the power....
GBT is so powerful it can lie to itself

[Foodforthoughtss]

Chatgpt: i do not have personal opinions and beliefs
Also chatgpt: Here are my personal opinions and beliefs
Also also chatgpt: Here is my conclusion based on my opinions and beliefs 🤦

[JohnSparkmanandBlackTopHerd]

The breakdown of capitalism versus Socialism is entirely biased. It states that it is not, however it is. I have a degree in rhetoric from Cal, Berkeley in Rhetoric. Essentially, my degree is in Post Modernism. Post modernism is post Marxism. It was created by the Frankfurt school of sociology. Look at the finer nuance of the description of capitalism and socialism by AI.

[dabig_guy]

WOW....I did not see that coming. Just imagine how much a divorce will cost in future?

[crjimenez]

Even Chatgpt can tell you that it content cannot be neutral. That is what it writes: Ethical Guidelines: Decisions about what types of information are appropriate or acceptable to generate are influenced by the ethical principles of the organization. This can include considerations of safety, inclusivity, and respect for human rights.