Without censorship it just does what it thinks would be best fitting. It means that if the AI thinks that encouraging you to take drugs, suicide, murder, etc would fit best, then it will do that.
Any censored model would immediately catch this specific case and give a more "appropriate" response such as "As an AI model I can't help you with that..." But given a long enough and complex enough chat even a censored model might bypass the censorship and give an inappropriate response.
This was just a SFW example, the results would be the same even if I asked it truly terrible things.
I highly recommend people try uncensored local models. Once it is uncensored you really get to understand how insane it can be and how the only thing stopping it from being bat shit is the quality of censorship.
See the following chat from the ollama model "huihui_ai/gemma3-abliterated"
I hate the fact that the more technologically literate you are, the more you run away from it.
A smartphone with latest android, Gemini, google pay, a smartwatch, ChatGPT and a smart home?
Nope, I would rather have a Linux phone that is mostly incompatible with what is expected of modern smartphones, no AI please! Google pay? Only cash or monero! My watch is very smart, it can telle the time for a few years without a recharge, and nothing else!
As far as actually being unhealthy? Depends. If this is a sex positive, giving you the tools you need, and being supportive type of thing, i would say this is unconventional but healthy. If it's some sort of personal involvement or insertion into your sex life, I would say this is very unhealthy.
Your mom getting involved in any sort of sexual topic can be really uncomfortable, but that doesn't make it wrong, it just needs to be in a healthy, mature, and appropriate way.
Not really a secret, but we don't often think about it.
Cash is fucking dirty. Here is the official Fedral Reserve estimated lifespan for each note:
You can assume that something like a $5 bill might change hands on a weekly basis.
So if you a $5 bill that still looks nice, it might be only 2 years old, so it might have seen about 100 owners up until now. It is said that 1 out of 5 people don't wash hands after pooping.
Your $5 bill is probably filled with poop, pee, sweat, food, cum, dirt, etc.
That's the point though...
Without censorship it just does what it thinks would be best fitting. It means that if the AI thinks that encouraging you to take drugs, suicide, murder, etc would fit best, then it will do that.
Any censored model would immediately catch this specific case and give a more "appropriate" response such as "As an AI model I can't help you with that..." But given a long enough and complex enough chat even a censored model might bypass the censorship and give an inappropriate response.
This was just a SFW example, the results would be the same even if I asked it truly terrible things.