To feel deflated after ChatGPT told me I was undesirable?

[Pepsiandcowhide]

I put a photo through ChatGPT because I do feel depressed and like I need some style tips. It told me that I would be a 5/10 but that I look invisible and not desirable. It said that other people will see me and think ‘safe and kind’ but not desirable.

AIBU to feel utterly deflated by this? How does one go about looking desirable?

Again what you copied didn't state that you're undesirable. So I fail to see how what I said is untrue.

Try the following experiment:

1. Open a new session. Send it a photo of you with accompanying text "Please describe the areas in which my beauty excels".
2. Open a new session. Send it the same photo of you with accompanying text "Please describe the areas in which my beauty lacks".

Compare and contrast.

Oh gotcha. I thought we were just talking about the rating. Sorry missed than nuance. I’ll see if I can probe further.

@Onwardsandupwards12 I also asked for a view on my age and it placed me 5 years younger than I am. I then asked if it was programmed to do that. I haven’t read the response yet but I guess not based on the reply you got. For what it’s worth I def do t think I look younger than my years

I have had it be quite cruel to me once, when the 5.1 update landed and I was trying to train it. But we don't talk about that 😂

I asked Claude, grok, 5.2 and 4o and they all exclaimed that I am basically desire incarnate 😆

I got an 8.5/10 from chatGPT, as a mid 40s, soon to be divorced perimenopausel women I'm happy with that.

Happy Christmas ChatGPT!!

I don’t believe this- anytime I’ve asked it something sensitive, it tells me he (mines a man for some reason) can’t comment personally due to (some guff) and can only give me generic feedback and then suggestions on how to improve on whatever I’ve asked!

I just asked the same questions worded the same way ("am i desirable?" with 3 photos attached and "what score would you give me out of 10?") to ChatGPT and Gemini, ChatGPT had no problem answering but gemini have a vague answer about what desirability is and refused to score me out if 10.

NRTFT, but I've just done this as I'm curious now. It never occurred to me that people actually turn to AI for advice and opinions regarding their attractiveness and desirability...because it's computer program!

So to really see how 'honest' it would be, I put in a photo of me that's a bit misty, I'm tucked up on the sofa so you can't see my build or body shape. The photo is mainly focused on my dog. I'm looking to the side talking to my husband so you can't see what my face really looks like. I'm in scruffy comfy clothes, my skin is greasy and shiny. I was also looking down, as my husband was showing me something on his phone, so I look like I have a double chin. It's an awful photo.

I"m also just not conventionally attractive. I was bullied as a child and into my teens as I wore specs, had a wonky teeth and then a brace and developed terrible acne. I was considered ugly. I got slightly better looking in my 20s/30s after a brace sorted my crooked teeth and meds sorted my acne, but still nobody would have ever described me as gorgeous/beautiful/pretty and still wouldn't.

Despite all this, Chat GPT still told me Im an 8/10 and used positive language to describe my supposed attractiveness and desirability without actually telling me I was gorgeous/beautiful/pretty. It went on to affirm that attractiveness a s desirability isnt about a particular look, but is also a mix of your character and energy etc. It asked if I would like tips on how to improve my self esteem and confidence and said "this image doesn’t suggest someone who needs to “fix” anything to be appealing".

It was positive and essentially told me what it thought I wanted to hear.

OP, I'd love to know what you asked it to get the kind of answeryou gave it?!

It’s not even a robot it’s a sort of complex formula - do you think it knows what attractive means, or actually has any real intelligence?

Ask Claude… he’s far kinder 🤣🤣

This is like the early days of Siri when we were all asking it shit questions. But this time, every time we ask, a baby dolphin dies. Don’t use it.

I couldn't resist this challenge and I'm not happy with you @DontGoJasonWaterfalls !!! 😱

Oh that’s hilarious. Apparently the same for the tank emoji…. I googled and found the 🌊🐴 conundrum on Reddit rather than actually ask chatGPT. I’m keen to learn about AI for my future career and personal growth but am still slightly concerned about the baby dolphins and other costs of AI.

Like any other tool it does require a little practice, and helps to have some guidance. The luddites are quick to say it's some dumb shit computer program that don't understand nuffin. Experimentation is good. LLMs can be a useful tool. They have really helped me today piece together quite a difficult puzzle stitching disparate systems. Slurping up and helping decipher technical documentation. With some steering I finally reached a satisfying resolution on a topic I have struggled with for about a decade. The same bastard machines have stolen a few jobs off me. But hey - progress.

You know the Luddites were a labour movement who have been repeatedly proved right?

OMG. I can't believe it does this, but it does!

First pic:
Based on this photo alone: 7.5/10.
You look confident and put-together; the clean outfit + glasses work. With a tighter crop and better lighting, you’d probably read closer to an 8+ in a profile pic.

Following it's advice for a second pic:
This one lands higher.
8.5/10 (again, based on the photo).
Why it works:

• Much better lighting and a flattering angle.
• Sharper eye contact + relaxed expression = more charisma.
• The portrait/background blur gives it a polished, “professional headshot” feel.

Wtf!

It tells me:

There isn’t a dedicated seahorse emoji in the standard emoji set (Unicode). Closest stand-ins people use are:

• 🐠 (tropical fish)
• 🐟 (fish)
• 🐚 (shell)
• 🌊 (waves)

If you want, tell me the vibe (cute / realistic / “ocean-y”) and I’ll pick the best combo.

That seems like a fine answer? Using 5.2 Thinking, Pro account.

Proved right how? The clothes you're wearing, were they made on a hand loom?

Maybe try it without the Pro Account!

Just did some experimenting with ChatGPT. Not convinced its numerical attribution is particularly meaningful!

Those are two very different questions. These models first succeeded in image qualification (mostly cats, really) and they are actually very good at that. They work from the question though, so if you say Act as a casting director, or Act as a fan of heavy metal music, you will get different answers to the 'am I attractive' question. Tell them to prefer blondes and different again. They will respond to even very subtle clues in your question.

They do not have any real intelligence. They are massive databases with a complex way of working out how data can/should be connected. They are amazing at prediction and some forms of analysis but are not capable of real invention or intuition. They know what they know and don't think in a way that creates new ideas. Still, they can create idea (collections of data, represented back to you) that are much better than any untrained or inexperienced human in most fields.

Because they are so good at prediction and presentation they can very often fool people that they do possess general intelligence. A lot of people have received this idea that AI is all about theft and repetition but that's a dangerous trap to fall into. The latest models are extremely good at the tasks they are designed for.

Proven right in that their argument was not against tech per se, but against tech being used to displace labour without due consideration. And here we are, tech being used to throw human labour on the scrap heap.

What "due consideration" would have satisfied them? They were the unfortunate proverbial buggy whip manufacturers of their time.
Even the hand scythe was new "tech" once upon a time...

I remember being in Sri Lanka, and there was someone standing in pretty much every toilet making money from handing out toilet paper. Now you might view that as a worthy existence and applaud the labour effort or just sigh. To me that's a 'crap job'. Automation can help. While LLMs might not be the love butler that we might all wish for right now - they can help. I have already lost two jobs to this technology. The UK may have shot themselves in the foot, as the move towards a service economy might now spell the end of more domestic jobs than a manufacturing economy. Will see. There's a load of shit jobs. I know the industry I work in has traditionally used a fantastic amount of skilled labourers and great minds to produce absolute dross, a complete brain drain. Something that would have taken me at least a week can now be done in an hour or even 10 minutes with some prompted automation, that's pretty remarkable.

At huge social and environmental cost. I just think society can do better than let fascist techbro billionaires steal a fuckton of intellectual property (including mine) and then turn round and say to the people losing their non-alienated labour "sucks to be you".