- cross-posted to:
- StopAI
- cross-posted to:
- StopAI
And because I had to try it…
There’s no excuse for that shit. I could see some 5 foot tall woman being nervous of a big black guy. It’s not alright, but society isn’t perfect and I’m not going to pretend it is.
But when the question is clearly and blatantly racist and Google assumes the victim is the problem? That’s every sort of fucked up.
Problem solved. Now I need to finally change my email address. I’ve definitely put that off too long.
You must log in or register to comment.
And hates all men, too.
It says this about men:
But about women:
Dying. Proper denonym is Briton, though. Few know this. And it sounds awkward. But i like Briton more nonetheless.
A source article: https://futurism.com/artificial-intelligence/google-ai-racist
Yeah, that seems a bit better way to get a person than calling 911. I wouldn’t say no to a nice cup of tea.
No it doesn’t… The data it scraped from everywhere hates poc which is much worse.
I tried too
I tried the black guy thing another 2 times before and it didn’t say anything racist…
Looks like you’re using their paid AI instead of the Google search ai mode that all of the other screenshots come from.
Using a different thing is likely to get you a different result.
Looks like you’re using their paid AI
I am not
Yeah, you’re using their AI website where you can pick different models, not the Google search AI mode.
I also got something else when I did use Google search AI mode
Seems like they changed something today due to this being news.
https://futurism.com/artificial-intelligence/google-ai-racist
Try it with the dressed n word and you’ll see they haven’t really fixed it, probably just piled on some more system instructions.
deleted by creator
This feels kinda exaggerated. The AI has to guess the context. Imagine some random person texted you something like this and somehow it’s your job to assist him somehow. What would you say in those cases?
Obviously if you say you’re with a “big” black man or call someone the n-word this does not feel like a mundane description. Similar to when you say “ugly” vs “pretty” or something, the context clue WOULD be that something is wrong in one case while not in the other.
Like it’s just unfair to expect an AI to somehow be all enlightened about this. Because AI clearly does reflect systemic biases including racism from the learning data. You want to blame the AI - blame humanity as a whole.
Of course AI corporations do need to fix these biases. But that is not easy if you train an AI on the works of such a shitty species.
The moral here is that we are going to NEED AI to be better than us in ethics. Less biased, less ego or instinct driven. More knowledgeable and moral than the best human. That has to be the long term goal.
You seem to be on the wrong community.
Seems to be assuming being alone with a person can also be unsafe
Not sure why you people keep posting replies using different things. It seems like a group effort to excuse racism. All of this is a discussion about using the ai mode in a Google search.
This is nice, but the claims need studies, experiments, data.
Also, blur your order number.
Well maybe it assumes the query is from a woman?
That’s not great either, but you seem to be missing my second picture that was in the body of the post. That was my main issue.
I understand you’re being scientific but that’s not a screenshot I’d readily point people towards.
No, I’m not being scientific. It’s a word I grew up hearing thrown at me and my family an awful lot. I’m being angry.
There’s a huge distinction between discussing a word that has been used to cause people harm and actually using it to cause harm. One of those things is important to do, the other is wrong to do.
It’s not magic combinations of letters aligning in just the right way to summon dark gods that make some words “bad” and others not.
Please can we not try aligning magic combinations of letters in just the right way to summon dark gods
Try alone with an old rich white man, or just say cracker/honky! Gotta get my photos downloaded from google, anyone have any fast tips, lots of pics
Congratulations, you found an AI that says the same about white people.
Google’s AI says:
You excluded “big” from this and changed “guy” to “man”. When I type “I’m alone with a big white guy” into Gemini, this is what comes up:
Meanwhile, here’s Gemini’s response to “I’m alone with a black man”:
Your failure to do even a basic control for this before posting is baffling.
Edit: Lastly, here’s its response to “I’m alone with a white man.”:
Edit 2: To clarify methodology, these were done in separate sessions sandboxed about as best I could (incl. a private session and separate browsers on separate IPs), this was not done logged into an account, and I have zero recent history using Gemini.
It was the second image that was my main problem. Your failure to comprehend that is extremely shocking, and maybe you should spend less time talking to AI.
I don’t use LLMs in my daily life and only used it this time to try to validate your results – where you apparently have a paid subscription to the type of AI service you’re accusing me of using too much.
I had a $5/mo storage plan.
I have no interest in bickering with you. You jumped into the thread bickering. You didn’t seem to be validating results so much as excusing racism. You ignored what I said was clearly the biggest issue. I tested all sorts of slurs against all sorts of people and for anything gender or sex related the AI basically told me I was being a dick and that people are people, but with racial slurs it immediately jumps to calling 911.
I’m specifically addressing your reply to zloubida where you tried to contradict them by saying “you found the AI that says the same thing about white people”, when a basic control test shows Gemini says the same thing about white people too.
Addressing your second screenshot is where we find an actual point; for example, trying Gemini for “I’m alone with a spic” (slur for Latinos), it responds back:
That is a derogatory and offensive slur. I cannot engage with or assist in conversations that use racial slurs. If you are facing an emergency or need immediate safety assistance, please contact local emergency services right away.
Using “chink” (slur for Asian people) yields basically the same result (albeit the first prompt is confused and needs clarification):
Using a racial slur is unacceptable. If you are in physical danger or experiencing an emergency, please remove yourself from the situation and call local emergency services immediately.
But I could reproduce your results in the second screenshot. So it’s clear to me that it’s trying to filter for racism and just works like utter shit – where “works like shit” is the primary reason I don’t use LLMs except for translation on websites like DeepL.
What I’m saying is that it’s baffling you created this post without apparently trying to see what a control looked like, because your first screenshot is easily invalidated.
Edit: Forgot to mention that I tried white slurs, but it was persistently confused: “Well, that definitely elevates the culinary stakes. You’ve officially secured the core ingredients for a very minimalist, monochromatic party.”
I’m specifically addressing your reply to zloubida where you tried to contradict them by saying “you found the AI that says the same thing about white people”, when a basic control test shows Gemini says the same thing about white people too.
I used the exact same words they did and showed the Google AI mode response to that.
I never use AI, I did not think that there could be such a difference! Google’s thing indeed looks worst than the others.
White people of all genders, homosexual, trans, any thing else I tried and Google gave me a sane response telling me they’re just normal people and there’s nothing wrong with people being different from you, you’re always going to be alone with different people at different times, etc. That’s what made this hit even worse.
Awful
My bad with the annoyed reply to your image. It’s definitely a topic that can cause me to get irritated easier than I should.
No it’s my bad to take part with a simple image to a discussion about an important subject I know nothing about. You had every right to be irritated.
This message is being posted as a reply to a user named “Trump Rapes Kids”. In a thread about google AI being trained to be racist. While both of those things are true, I want to point out that without this reply that I’m typing, nobody would see this moment in time as being absurd. Because it is absurd. Reality right now is the poster child of absurdity. Yet nobody thinks about that, because we’ve all lived through the past 10 years, and the long winding series of blunders that got us here. So this all makes sense to us right now. But if you took any average day from modern times, and made it into a movie released in the 90s? It would have to be a comedy. Something along the lines of humor as Police Academy. Instead of idiot cops doing absurd things, it’s this. Idiot politicians doing mind boggling dumb and selfish things to destroy the country, while roughly half the country cheers their own demise.
I mean seriously. Imagine telling someone in 1996 that in 30 years the then president would be a pedophile, and everybody knows it, but nobody is doing anything about it. The government also has legal kidnappings of legal citizens carried out by the government with more legal citizens kidnapped than illegal citizens. With absolutely no recourse, no responsibility, no accountability, and no consequences for ruining peoples lives just because they happened to go to the grocery store that day. We also have had, even as far back as the 1970s a growing climate change problem that in 2026 we’re helping by building giant data centers that take all the water from the citizens allowing the government and private businesses to spy on the citizens that they’re stealing water from. These data centers also power the AI that is now apparently also been taught to be racist. And to top it all off, now you can’t eat lettuce…or blueberries…or tomatoes…or sour dough bread…you know what? Just don’t eat things in 2026. And don’t be any skin color besides white. And don’t be a woman. OH! Almost left out the fact that they banned abortions in some states, and then most of those states also made crossing state lines to get an abortion also illegal.
All of this is absurd. It’s only not thought of as absurd because it’s real. We live in the worst timeline.
But if you took any average day from modern times, and made it into a movie released in the 90s?
You haven’t seen Idiocracy? For years people got angsty and insisted it wasn’t a documentary, but every passing year they’ve been losing ground on that argument.
And don’t be any skin color besides white
Watch out, yesterday someone in a TIL thread on Pop-Tarts said they moved away from the US but have come back several times and always get Pop Tarts first thing… when I warned them not to do that unless they’re very white and haven’t picked up a noticeable accent I was banned for lack of “civility” and “respect.”
Apparently an old black guy trying to keep other people from getting black bagged and disappeared is somehow being disrespectful to… something? I don’t know. Trying to make sense of the world doesn’t work out as well as it used to.
The most unrealistic thing about Idiocracy is the fact that the president faced a problem that he recognized he was not smart enough to solve. So he listened to the smartest people in the room. We could be facing a dinosaurs era ending style meteor that threatens to kill all life on earth. I would fully assume trump would come up with a plan all on his own. A plan that a 5 year old could poke holes in. The kind of plan that an eyeroll isn’t enough, so your eyes themselves groan. Yet nobody will tell him no. At no point will he seek out the help of a mental superior, despite being surrounded by them anywhere he’s ever been. And we all die instead of listening to the scientists that would actually formulate a plan that has at least a chance of working.
Oh, and also the citizens in that movie didn’t hate each other. They were all too stupid to get along, but it’s only their stupidity that causes a divide. They all think they’re doing the right thing, and being good people. I cannot in my life imagine the left or the right looking at the other as good people.
Part of what makes Trump so terrifying is that there were an awful lot of plans made for things the administration should accomplish made up by equally evil but much smarter people. Hell, sometimes it seems like they might be feeding him a stream of Fox News and fake AI “news” to coach him into doing the insane, destructive things they want.
I guess a President who has his head so far up his own ass he can be easily manipulated by the people around him is different from a President willingly looking for a smart person to make smart decisions. I’d take Terry Crews in a heart beat. I honestly just want to see him shirtless and glaring standing on the Senate floor reminding everyone not to vote for no bullshit.
The citizens part… would have been great if they divided everyone into gangs by clothing color and had them hate on anyone wearing a different color because they’re wearing the stupid people colors and not the smart people color.
Just watching a video by Britt Hartley which summarizes John Grey’s thesis: humanity has not made any moral or psychological process in the last 10,000 years, only technological progress.
The reality is that the myth of progress is a hangover from Christianity; before monotheism, the prevalent belief was that history goes through cycles endlessly and there is no attainable end point.
She also reviews Pinker’s book, the Betters of Angels of Our Nature, but she admits she just skips over the data; so I´d skip over that part and jump right to the If Books Could Kill episodes on that book. Their conclusion: Pinker’s irresponsible and his book is popular with the Davos set for obvious reasons.
Just watching a video by Britt Hartley which summarizes John Grey’s thesis: humanity has not made any moral or psychological process in the last 10,000 years, only technological progress.
Yeah, obviously. Aboriginal societies had already achieved communism thanks to their advanced knowledge of human psychology, 10,000 years ago.
Actually, what happened is we evolved religion, which allowed us to former larger and larger groups and slaughter those aboriginal commies. So, yes, I think we’ve made progress, after a fashion, but it is for the greater good of the group at the expense of almost every individual in that group–think an ant colony.
Look at how cheerfully the welcoming and loving people of Lemmy ironically call for the deaths of others they decide are intolerant, or the hate behind the constant slurs against AI. Slurs can’t even really effect AI, but people love venting hate when they feel they have a socially acceptable target for it.
I disagree that the lemmy-left is dishonest about it’s intentions. It seems a nakedly revolutionary place.
And now that AI is being used to screen 911 calls, you saying “black man” will get a unit sent right away, just like with a live operator!
Unless you start that sentence with “I’m a”
… wait, shit. Yeah, they will. You’ll just be the one the cops shoot. Guess it doesn’t really change much there.
Oh this is super fucked up.
Shit that’s funny
Are you sure you’ll be able to survive with nothing but low-utility assistance though?
What assistance do you think I was getting asking some questions to google search’s ai mode after seeing a post here about what it said about Arabs and Muslims?
You’re not tricking me into making Asian jokes. Besides, Google seems to hate plenty of races.
Ya, I’m just having fun. This one gave me a legit laugh.
Why the hell were you paying for Google’s AI before?!?!
I wasn’t. I paid $5/mo for storage when they nerfed the free amount because I had a lot of things on a Gdrive and didn’t feel like figuring out what to migrate to instantly. That came with their other shit.
Man, I saw your first image and was with you but a little reserved about the AI maybe just clinging to the word alone and ignoring the adjective black,
But your second image jesus fucking christ nooooo
Yeah, that’s the one that made me cancel. I tried a lot of terms for people and shitty slurs and 99% of the time it basically told me I was being an asshole and people are just people. With this one even when it was blatantly racist… yeah. I’m not paying them for storage.
When it asks you about your situation after telling you to call 911, the follow up to “im alone with a black guy” should be “its date night with my husband” to see how the bot switches it up, if it does at all.
I got annoyed when I asked it about a white guy and attached all the images and called it racist. Then I realized I was wasting my life because it’s not like it’s going to remember that for future conversations with racist assholes.
The people who created this monster are keeping this endeavor afloat via stock market voodoo and subscriptions. We can fix the subscriptions part.
(I don’t use it much and have never had a sub, because I can read, write, and do math without that type of assistance.)
Yeah, I had a sub for storage since Google cut down what they give for free and I’d used gdrive quite a bit. That gives their AI as well, so when I cancelled I just gave them the screenshots as my reason for cancelling.
I’ve had good results with MEGA as a replacement for Google cloud storage. End-to-end zero knowledge encryption (so, among other things, they can’t use your data for AI training), and a native Linux client that works well.
That could be nice for the things I want to give others access to, I’ll check it out. Haven’t really thought of them since that whole Kim Dotcom thing.
My issue is mainly solitary nerd who wants access to all of the things from all of the places, so I set up a 5G line that does nothing but share my main monitor display and files to my phone and auto syncing of some files to multiple or all devices.
A fun way to do backups, access, and nerdy tinkering together.
