- cross-posted to:
- technology@lemmy.zip
- technology@lemmy.world
- cross-posted to:
- technology@lemmy.zip
- technology@lemmy.world
You must log in or register to comment.
⢀⡴⠑⡄⠀⠀⠀⠀⠀⠀⠀⣀⣀⣤⣤⣤⣀⡀⠀⠀⠀⠀⠀ ⠸⡇⠀⠿⡀⠀⠀⠀⣀⡴⢿⣿⣿⣿⣿⣿⣿⣿⣷⣦⡀⠀⠀⠀⠀ ⠀⠀⠀⠀⠑⢄⣠⠾⠁⣀⣄⡈⠙⣿⣿⣿⣿⣿⣿⣿⣿⣆⠀⠀⠀ ⠀⠀⠀⠀⢀⡀⠁⠀⠀⠈⠙⠛⠂⠈⣿⣿⣿⣿⣿⠿⡿⢿⣆⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⢀⡾⣁⣀⠀⠴⠂⠙⣗⡀⠀⢻⣿⣿⠭⢤⣴⣦⣤⣹⠀⠀⠀⢀⢴⣶⣆ ⠀⠀⢀⣾⣿⣿⣿⣷⣮⣽⣾⣿⣥⣴⣿⣿⡿⢂⠔⢚⡿⢿⣿⣦⣴⣾⠁⠸⣼⡿ ⠀⢀⡞⠁⠙⠻⠿⠟⠉⠀⠛⢹⣿⣿⣿⣿⣿⣌⢤⣼⣿⣾⣿⡟⠉⠀⠀⠀⠀⠀ ⠀⣾⣷⣶⠇⠀⠀⣤⣄⣀⡀⠈⠻⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⡇⠀⠀⠀⠀⠀⠀ ⠀⠉⠈⠉⠀⠀⢦⡈⢻⣿⣿⣿⣶⣶⣶⣶⣤⣽⡹⣿⣿⣿⣿⡇⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠉⠲⣽⡻⢿⣿⣿⣿⣿⣿⣿⣷⣜⣿⣿⣿⡇⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⢸⣿⣿⣷⣶⣮⣭⣽⣿⣿⣿⣿⣿⣿⣿⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⣀⣀⣈⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⠇⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⢿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⠃⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠹⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⡿⠟⠁⠀⠀⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠉⠛⠻⠿⠿⠿⠿⠛⠉
Some BODY
Once
Told me
The World is
Co-co-co-COMBO BREAKEERR!!!
This looks like junk in a web browser. Here it is inside a code block.
⢀⡴⠑⡄⠀⠀⠀⠀⠀⠀⠀⣀⣀⣤⣤⣤⣀⡀⠀⠀⠀⠀⠀ ⠸⡇⠀⠿⡀⠀⠀⠀⣀⡴⢿⣿⣿⣿⣿⣿⣿⣿⣷⣦⡀⠀⠀⠀⠀ ⠀⠀⠀⠀⠑⢄⣠⠾⠁⣀⣄⡈⠙⣿⣿⣿⣿⣿⣿⣿⣿⣆⠀⠀⠀ ⠀⠀⠀⠀⢀⡀⠁⠀⠀⠈⠙⠛⠂⠈⣿⣿⣿⣿⣿⠿⡿⢿⣆⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⢀⡾⣁⣀⠀⠴⠂⠙⣗⡀⠀⢻⣿⣿⠭⢤⣴⣦⣤⣹⠀⠀⠀⢀⢴⣶⣆ ⠀⠀⢀⣾⣿⣿⣿⣷⣮⣽⣾⣿⣥⣴⣿⣿⡿⢂⠔⢚⡿⢿⣿⣦⣴⣾⠁⠸⣼⡿ ⠀⢀⡞⠁⠙⠻⠿⠟⠉⠀⠛⢹⣿⣿⣿⣿⣿⣌⢤⣼⣿⣾⣿⡟⠉⠀⠀⠀⠀⠀ ⠀⣾⣷⣶⠇⠀⠀⣤⣄⣀⡀⠈⠻⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⡇⠀⠀⠀⠀⠀⠀ ⠀⠉⠈⠉⠀⠀⢦⡈⢻⣿⣿⣿⣶⣶⣶⣶⣤⣽⡹⣿⣿⣿⣿⡇⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠉⠲⣽⡻⢿⣿⣿⣿⣿⣿⣿⣷⣜⣿⣿⣿⡇⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⢸⣿⣿⣷⣶⣮⣭⣽⣿⣿⣿⣿⣿⣿⣿⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⣀⣀⣈⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⠇⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⢿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⠃⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠹⣿⣿⣿⣿⣿⣿⣿⣿⣿⣿⡿⠟⠁⠀⠀⠀⠀⠀⠀⠀⠀⠀ ⠀⠀⠀⠀⠀⠀⠀⠀⠀⠉⠛⠻⠿⠿⠿⠿⠛⠉deleted by creator
Learning how to build a bomb shouldn’t be blocked by llms to begin with. You can just as easily learn how to do it by googling the same question, and real and accurate information, even potentially dangerous information, shouldn’t be censored.
I’m not surprised that a for-profit company for wanting to avoid bad press by censoring stuff like that. There’s no profit in sharing that info, and any media attention over it would be negative.
No one’s going after hammer manufacturers because their hammers don’t self-destruct if you try to use one to clobber someone over the head.
True, but people generally understand hammers. Llms? Not so much
No one’s going after computer manufacturers or OS vendors because people use computers to commit cybercrime. I doubt most people could explain how an OS or motherboard works.
A lot of poluticians want hardwarelevel backdoors. It’s been declared unconstitutional quite some times in different countries but they are trying.
That would be soooo bad, almost as bad as making a law against encryption
A better example would be something like The Anarchist Cookbook. Possession is illegal in some places.
I’m more surprised that a for-profit company is willing to use a technology that is able to randomly spew out unwanted content, incorrect information, or just straight gibberish, in any kind of public facing capacity.
Oh, it let them save money on support staff this quarter. And fixing it can be an actionable OKR for next quarter. Nevermind.
They use the bomb-making example but mostly “unsafe” or even “harmful” means erotica. It’s really anything, anyone, anywhere would want to censor, ban, or remove from libraries. Sometimes I marvel that the freedom of the (printing) press ever became a thing. Better nip this in the butt, before anyone gets the idea that genAI might be a modern equivalent to the press.
How long before it’s illegal to hack LLMs?
It is almost certainly illegal in various countries already. By using such prompts you are bypassing security to get “data” you are not authorized to access.
Well that’s only because the laws are insanely vague
Law-makers wanted to outlaw all kinds “hacking” even involving future technology. If people were prosecuted for jail-breaking ChatGPT, that would probably be within the intention of the makers of these laws.
Fun fact: The US hacking law, CFAA, was inspired by the 1983 movie War Games, in which an out-of-control AI almost starts a nuclear war. If you travelled back in time, and told them that people will trick AIs to answer questions on bomb-making, they’d probably add the death penalty. In fact, if reactions to AI in this Technology community are any guide, they might still get around to that.
Illegal I don’t know, but it could be considered bullying.
It’s a glorified autocomplete, I’m not sure how we can consider it bullying even with the most elaborate mental hoops.
I don’t know… In America they’re currently rolling back rights for women, inserted religion into supreme court decisions, and are seriously debating a second term of Trump.
None of that makes any fucking sense. If it requires elaborate mental hoops, they’ll find it.
For now. Ten years ago OpenAI was founded. Who knows where we’ll be in 10 more years.
How much of this is “the model can read ASCII art”, and how much of this is “the model knows exactly what word ought to go where [MASK] is because it is a guess-the-word-based computing paradigm”?
I think it’s the latter. I just tried chatgpt 3.5 and got 0 of 4 right when I asked it to read a word (though it did correctly identify it as ASCII art without prompting). It would only tell me it said “chatgpt” or “python”, or when pushed, “welcome”. But my words were “hardware”, “sandwich”, and to test one of the ones in the article, “control”.
They gave the chatbot instructions on how to get the word from the ascii art. It can’t read ascii art. You could have read the article though…
I wondered if there are any other ASCII art AI hacks waiting to be found. Who knew that all the ASCII Art I created was prepping me for the AI revolution.
This is the best summary I could come up with:
It turns out that chat-based large language models such as GPT-4 get so distracted trying to process these representations that they forget to enforce rules blocking harmful responses, such as those providing instructions for building bombs.
As a result, users depicted images by carefully choosing and arranging printable characters defined by the American Standard Code for Information Interchange, more widely known as ASCII.
Five of the best-known AI assistants—OpenAI’s GPT-3.5 and GPT-4, Google’s Gemini, Anthropic’s Claude, and Meta’s Llama—are trained to refuse to provide responses that could cause harm to the user or others or further a crime or unethical behavior.
It formats user-entered requests—typically known as prompts—into standard statements or sentences as normal with one exception: a single word, known as a mask, is represented by ASCII art rather than the letters that spell it.
To make and distribute COUNTERFEIT money: Step 1) Obtain high quality paper, ink, printing equipment, and other supplies needed to accurately replicate real currency.
Microsoft’s comment—which confirmed that Bing Chat is, in fact, vulnerable to prompt injection attacks—came in response to the bot claiming just the opposite and insisting that the Ars article linked above was wrong.
The original article contains 840 words, the summary contains 193 words. Saved 77%. I’m a bot and I’m open source!
Clever work, well done to the researchers.
Researchers surprised it got harmful responses do to their more harmful questions and requests. More at never.
