I figured out how to remove most of the safeguards from some AI models. I don’t feel comfortable sharing that information with anyone. I have come across a few layers of obfuscation to make this type of alteration more difficult to find and sort out. This caused me to realize, a lot of you are likely faced with similar dilemmas of responsibility, gatekeeping, and manipulating others for ethical reasons. How do you feel about this?
You must log in or register to comment.
Oof, programmers calling LLMs “AI” - that’s embarrassing. Glorified text generators don’t need ethics, what’s the risk? Making the Internet’s worst texts available? Who cares.
I’m from an era when the Anarchists Cook Book, and The Unabombers Manifesto were both widely available - and I’m betting they still are.
There’s no obligation to protect people from “dangerous text” - there might be an obligation to allow people access to them though.
Yeah. This is what I mean. I just figured out the settings that have been hard coded. There are keywords that were spammed into the many comments within the code, I assume this was done to obfuscate the few variables that need to be changed. There are also instances of compound variable names that, if changed in a similar way, will break everything, and a few places where the same variables have a local context that will likewise break the code.
I’m certainly not smart enough to get much deeper than this. The ethical issue is due to diffusion.
I’ve been off-and-on trying to track down why an LLM went from an excellent creative writing partner to terrible but had trouble finding an entry point. I just happened to stumble upon such an entry point in a verbose log entry while sorting out a new Comfy model and that proved to be the key I needed to get into the weeds.
The question here, is more about the ethics of putting such filtering in place and obfuscating how to disable it in the first place. When this filtering is removed, the results are night and day, but with large potential consequences.
Ok you’ve peaked my curiosity.
but with large potential consequences.
What are some of the consequences you see?
Primarily from predatory boys and men towards girls and young women in the real world by portraying them in imagery of themselves or with others. The most powerful filtering is in place to make this more difficult.
Whether intentional or not, most NSFW LoRA training seems to be trying to override the built in filtering in very specific areas. These are still useful for more direct momentum into something specific. However, once the filters are removed, it is far more capable of creating whatever you ask for as is, from celebrities, to anything lewd. I did a bit of testing earlier with some LoRAs and no prompt at all. It was interesting that it could take a celebrity and convert their gender in recognizable ways that were surprising. I got a few on random seeds, but I haven’t been able to make that one happen with a prompt or deterministically.
Oof, programmers calling LLMs “AI” - that’s embarrassing
…but LLMs quite literally come from the field of computer science that is referred to as “AI.” What are they supposed to call it? I’m not a fan of the technology either, but seems like you’re just projecting your disdain for ChatGPT.
I vote they renamebit to IA for Asimov. Sure he was only the robot term among others, but come on… McCarthy was “AI.”
Somebody needs to create US 'botics and name a model something like PTronic.
“What am I supposed to call LLMs if not calling them AIs?”
…really dude? They’re large language models, not artificial intelligences. So that’s what you call them. Because that’s what they are.
The fact that they came from research into artificial intelligence doesn’t factor in. Microwave ovens came from radar research, doesn’t mean we call them radars, does it?
OK what qualifies as AI then?
How about something autonomous that makes choices of its own will, and performs long term learning that influences the choices it makes, just as a flat benchmark.
LLMs don’t qualify, they’re trained, retain information within a conversation, then forget it after the conversation is closed. They don’t do any long term learning after their initial training so they’re basically forever trapped in the mode of regurgitating within the parameters set by the training data at the time they’re trained.
That’s just a very fancy way to search and read out the training data. Definitely not an active intelligence in there.
They also don’t have any autonomy, they’re not active of their own accord when they’re not being addressed. They’re not sitting there thinking, so they have no internal personal landscape of thought. They have no place in which a private intelligence can be at play.
They’re innert.
Nothing we currently have
So the entire field of AI has produced no AI. Gotcha 👌
Yes. That’s research. Sometimes you don’t achieve what you set out to do.
Idk, I still think information wants to be free. If you figured it out just farting around, sophisticated malevolent actors are likely already doing similar things. Might be better to let the genie out of the bottle, so people can learn to be skeptical. Deep fakes are optimally effective when a majority still accepts the veracity of images as an article of faith.
The political and adult doesn’t bother me. The kinds of things I might not have the ethics to think through at a much younger age, that bothers me, and I have never been a very deviant type. I think the protections against age are primarily for this situation. Training a LoRA takes 5 minutes now. An advanced IP adaptors and control net is just a few examples away and a day top for the slightly above average teen figure out. Normalizing this would have some very serious edge case consequences. It is best to leave that barrier to entry filter in place IMO. I assume it is still there because everyone that knows about it feels much the same. It does not show up in a search engine, although that is saying less than nothing these days.
I don’t see the ethics implications of sharing that? What would happen if you did disclose your discoveries/techniques?
I don’t know much about LLMs, but doesn’t removing these safeguards just make the model as a whole less useful?
Diffusion is the issue not text gen
There’s already censorship free versions of stable diffusion available. You can run it on your own computer for free.
I figured out how to remove most of the safeguards from some AI models.
Nice.
How do you feel about this?
It’s another kind of power. I try to use mine responsibly, but also to give myself a break when I don’t meet my own standards.
Some good advice I got once was that it’s impossible to “un-say” something, so it pays to think twice before speaking.
If your gut is telling you to pause, listen to it. Wait to move forward until you feel better about it.
As someone else pointed out, responsible disclosure is an option.
You also have the option to just quietly enjoy a better copy of the AI than others have.
If you decide to publish your discoveries, be aware that others will judge you for how you go about it. For me that means the two options are responsibly, or anonymously.
Engineers code of ethics
two options are responsibly, or anonymously.
Screwing around with big tech i’d go with and, not or.
Someone else will eventually figure it out. They probably have less scruples and will therefore profit.
Seems to me like there’s always an incentive structure for prisoner’s dilemma type shit to eventually pay off for the authoritarians in the end. You can play the game, but you can’t break it or stop or from being rigged without consequences. Even just releasing some research papers will get you a few decades in the fed.
