this post was submitted on 11 Mar 2026
59 points (94.0% liked)

Technology

82488 readers
3787 users here now

This is a most excellent place for technology news and articles.


Our Rules


  1. Follow the lemmy.world rules.
  2. Only tech related news or articles.
  3. Be excellent to each other!
  4. Mod approved content bots can post up to 10 articles per day.
  5. Threads asking for personal tech support may be deleted.
  6. Politics threads may be removed.
  7. No memes allowed as posts, OK to post as comments.
  8. Only approved bots from the list below, this includes using AI responses and summaries. To ask if your bot can be added please contact a mod.
  9. Check for duplicates before posting, duplicates may be removed
  10. Accounts 7 days and younger will have their posts automatically removed.

Approved Bots


founded 2 years ago
MODERATORS
 

At a glance, the passwords the LLMs created looked secure, much like those that a password generator might spit out. But that’s exactly where the problems arose: Although the AI-generated passwords appeared to be complex and safe to use for securing online accounts, they were actually quite predictable upon closer inspection.

All three LLMs exhibited clearly identifiable patterns in how they created these passwords. These patterns included repeated character strings, predictable password structure, frequent reuse of similar characters, clear biases toward certain numbers and letters, and even duplicate passwords in some cases. Although the AI-generated passwords looked random, they really weren’t. This could easily create a false sense of security if you were to use these predictable passwords for your online accounts.

top 49 comments
sorted by: hot top controversial new old

AI generated password? What kind of brain rot idea is that, god damn. Do these people need to ask AI to breathe for them?

I can also generate a password for them while running 50 toasters and flushing the toilet a bunch if they want.

[–] Artisian@lemmy.world 1 points 51 minutes ago

Just noticing: there's 0 evidence in article that anyone is doing this. I just don't buy that this is happening enough to matter. Interesting as interpretability research at best

[–] HubertManne@piefed.social 18 points 3 hours ago (1 children)

why. dafuq. would someone use an llm to make a password.

[–] dhork@lemmy.world 4 points 2 hours ago

Because the techbros are pushing AI (really LLMs, but that is too many letters) for everything to justify their insane stock valuations

[–] cyberpunk007@lemmy.ca 7 points 2 hours ago

"here's what to use instead"

...the same thing we always have been? Keepass or other password managers with randomly generated passwords?

[–] Randomgal@lemmy.ca 7 points 3 hours ago (1 children)

My brother you can literally slam your face on the keyboard to get a password. Why would you ask an llm' to do it for you?

[–] Dindonmasker@sh.itjust.works 2 points 3 hours ago (1 children)

I'm pretty sure there's an algorithm somewhere to solve for face to keyboard password generation.

[–] FauxLiving@lemmy.world 2 points 2 hours ago

Let me ask ~~ChatGPT~~ Claude

[–] cheese_greater@lemmy.world 84 points 7 hours ago* (last edited 7 hours ago) (7 children)

Why does everything have to be LLM, like why cant things sometimes just be algorithmically generated like my AI-free password manager does?

I wonder if there's AI-powered password managers on the market now lol. If so and if customers, I would mirror Zuck's snide "thuh dumb fucks" setiment when people trusted him

ZXCVBN algorithm has never failed me

[–] Naich@lemmings.world 27 points 7 hours ago (1 children)

I remember the days before hardware random number generators, when pseudo random number generators caused fuck ups by not actually being random. Now people are re-inventing the stupid way of doing it with AI.

[–] cheese_greater@lemmy.world 4 points 7 hours ago

It blows my mind hackers still found ways in with just that one fuckup

[–] artyom@piefed.social 8 points 5 hours ago* (last edited 5 hours ago) (1 children)

There's AI-powered everything now for no reason other than because when they slap those two letters on it, their stocks go up. A lot of times there's not even any AI involved.

I remember CES 2022 when derb8uer found these "AI PC cases" and when he asked about why, they said "because you can put an AI PC in it". And that's when I knew it was a bubble. Because that's the kind of shit that happened with dotcom. Companies slapping "dotcom" on their names and their stocks would skyrocket for no reason.

[–] cheese_greater@lemmy.world 2 points 5 hours ago

Stox go brrrrr

[–] BrianTheeBiscuiteer@lemmy.world 4 points 4 hours ago

I always say I'm not interested in asking an LLM to add 2 + 2 for me. Of course my managers are always pushing us to use LLMs and coworkers keep suggesting we replace efficient, testable, and consistent processes with AI. If you're going to use it at least think of scenarios that are hard to code for and it would take you at least 10 min to solve.

[–] BrainBow65@lemmy.world 14 points 6 hours ago (2 children)

It's worse than just using a new tool though. By definition LLMs use the statistically most likely option (with minor variation for flavor). People are literally asking a statistics bot what password can I use that is the most likely?

[–] grue@lemmy.world 1 points 2 hours ago

That's a good point: it's not just that LLMs fail to give you an optimal password, it's that they're inherently designed to give you a pessimal one.

[–] cheese_greater@lemmy.world 5 points 6 hours ago* (last edited 6 hours ago)

I was gonna say that. They gonna give you a recycled pattern

I mean, like, what is an effective/"good" password if not an unrecycled non-pattern??

[–] Madrigal@lemmy.world 7 points 7 hours ago

But…but Magic Box™

[–] Septimaeus@infosec.pub 0 points 3 hours ago* (last edited 3 hours ago) (1 children)

While “routing” of prompts between specialist models and traditional APIs does offer more efficient and reliable outputs, it also

A. requires more elbow grease than one massive generalist model, and

B. doesn’t help you avoid paying API licensing fees by obfuscating their outputs into the blackbox weights of your proprietary model

[–] cheese_greater@lemmy.world 3 points 3 hours ago (1 children)

It just seems like a password manager is an infinitely better tool. Like to the point i severely judge and question the sense of someone who seeks to reinvent the wheel for something so crucial

[–] Septimaeus@infosec.pub 2 points 3 hours ago (1 children)

Oh for sure. I was just answering in case you wanted to know.

Future versions of AI tools will likely use routing to do a variety of tasks more efficiently and accurately. But until then? It’s a jinn to those who would wish away every single chore they lack the patience to comprehend including, unfortunately, information security.

[–] cheese_greater@lemmy.world 1 points 3 hours ago

Its honestly crazy to me. I have it down to such an art that Incouldnt imagine giving up my unilateral control or sole custody of that data

[–] Eyekaytee@aussie.zone -1 points 6 hours ago (1 children)

Why does everything have to be LLM

Not everything is, it's just Lemmy is obsessed with them so you're going to see it come up a lot

[–] orclev@lemmy.world 4 points 6 hours ago

Unfortunately it's not just Lemmy. Corporations are obsessed with LLMs as are a lot of the people working for them. I daily get messages and meetings about LLMs in one way or another and it drives me nuts because it keeps interrupting actual work. In the span of like two months it went from something that was mentioned maybe once in the last year to be 90% of what everyone is talking about.

[–] asudox@lemmy.asudox.dev 43 points 7 hours ago (4 children)

which dumbass lets LLMs generate their passwords?

[–] binarytobis@lemmy.world 1 points 1 hour ago

If someone I knew told me they asked an LLM to generate their password, I would no longer talk to that person.

[–] Deestan@lemmy.world 1 points 2 hours ago

What do you mean. Why would it be a problem that my password is the most statistically likely sequence of characters in the world? And that the password is stored in plaintext in a chat log? And used for training LLMs on password conversations?

[–] Deceptichum@quokk.au 19 points 7 hours ago (2 children)

I just set my password to hunter2, no one will ever figure it out.

[–] dhork@lemmy.world 1 points 2 hours ago

I just use the same combination that I have on my luggage

[–] popekingjoe@lemmy.world 16 points 7 hours ago (1 children)

What is that? All I see is *******.

[–] crandlecan@mander.xyz 6 points 6 hours ago (1 children)

Let me test that

JsfJhdYhb57'j4++€€djF

[–] crandlecan@mander.xyz 6 points 6 hours ago (1 children)
[–] whyrat@lemmy.world 5 points 5 hours ago (1 children)

Don't put a quote mark in your password! I learned the hard way with password'drop table users;'

/S

[–] jcorvera@quokk.au 2 points 2 hours ago

Wait, was this because you too had a student named "Robert'); DROP TABLE users;--"?

[–] crandlecan@mander.xyz 2 points 7 hours ago (1 children)
[–] Triumph@fedia.io 2 points 7 hours ago (1 children)
[–] crandlecan@mander.xyz 4 points 6 hours ago

That was uncalled for 😞

[–] yggstyle@lemmy.world 28 points 6 hours ago

Why would anyone... Who the fuck is using glorified autocomplete to generate a random unique password?

Anyone that stupid deserves the breach and any company that integrates that moronic idea deserves the blowback.

[–] Diplomjodler3@lemmy.world 8 points 5 hours ago (2 children)

It takes a five line Python script to generate a secure password. But no, let's waste a gigantic amount of energy to have a slop machine do it very badly.

[–] ShinkanTrain@lemmy.ml 1 points 2 hours ago

Even less lines if you're ok with it generating the same secure password every time

[–] kayzeekayzee@lemmy.blahaj.zone 1 points 2 hours ago (1 children)
[–] kayzeekayzee@lemmy.blahaj.zone 1 points 2 hours ago* (last edited 2 hours ago) (1 children)
import random
password_length=10
letters = "abcdefghijklmnopqrstuvwxyz0123456789-&@$?!*#%"
print(''.join(random.sample(letters+letters+letters.upper()+letters.upper(), password_length)))

If you inline password_length and letters, you can bring it down to 2 lines

[–] Diplomjodler3@lemmy.world 0 points 2 hours ago (1 children)

Sure. The five lines are just for readability.

golfing is fun! I do everything in a single line

[–] flamingo_pinyata@sopuli.xyz 14 points 7 hours ago

LLMs are predictive text machines. Focus on "predictive". Of course they will not output random text.

Note: not fully deterministic though - they need (pseudo)randomness at few critical points to be good

[–] Th4tGuyII@fedia.io 2 points 6 hours ago

How did we manage to move almost completely away from Pseudo-random password generation, only to come straight back to it via LLMs again.

People will really just shove these things into anything like they're a bloody panacea, when really they're just extremely well-trained parrots.

[–] andybytes@programming.dev 1 points 5 hours ago* (last edited 5 hours ago)

AI is a scam bruh... snap snap wake up unit 242... clap AI clap IS clap A clap BUZZWORD... like the biggest techbro scam yet. It will be ok you are just watching the boat sink as the economy shrinks like ol donald dumpy dick. You better tell him he is pretty or he will end the world. Waymo admits that its autopilot is often just guys from the Philippines

**Crapitalism tricked your arse again. **

[–] ptu@sopuli.xyz 1 points 7 hours ago

Is there a link to the actual study?

[–] andybytes@programming.dev 0 points 5 hours ago* (last edited 5 hours ago)

AI is like going into the kitchen and telling the chef how to cook. A time waster, fear monger tactic by scared middle mangers afraid they will get replaced if they don't adopt and push it. When dumb dumb yankkkee land is done fucking around they are still gonna need people, they just want to pay you low wages. You are now the bottom bitch. gotta keep these workers in their place by running off of fumes. Reinvent the wheel as a square and beat me silly with it.

AI is perpetual motion machine on the planet WASP-127b .....which is no vacum. You worry about the memes as the crapitalist crush your dreams. A debt you will never pay with a shrinking economy. AI did not take your job... your job doesn't exist... the epstien class takes a weee in your cornflakes thats why they taste like piss. A magic trick... your fear of accepting slop... It was already slop and your job was already dumb. Circling the drain as the rich play their games. Self reinforcing Exponential feed back loops on a road to no where, on tracks coming to an end. AI is a sloppy dumby drag net that paints with broad strokes... lower the bar cut the coke, the people at the top are stupid jokes. You are bound by thin treads of silk...