[a / b / c / d / e / f / g / gif / h / hr / k / m / o / p / s / t / u / v / vg / vm / vmg / vr / vrpg / vst / w / wg] [i / ic] [r9k / s4s / vip] [cm / hm / lgbt / y] [3 / aco / adv / an / bant / biz / cgl / ck / co / diy / fa / fit / gd / hc / his / int / jp / lit / mlp / mu / n / news / out / po / pol / pw / qst / sci / soc / sp / tg / toy / trv / tv / vp / vt / wsg / wsr / x / xs] [Settings] [Search] [Mobile] [Home]
Board
▼ Settings Mobile Home
/pol/ - Politically Incorrect

Name
Options
Comment
Verification
4chan Pass users can bypass this verification. [Learn More] [Login]
Flag
File
  • Please read the Rules and FAQ before posting.

08/21/20New boards added: /vrpg/, /vmg/, /vst/ and /vm/
05/04/17New trial board added: /bant/ - International/Random
10/04/16New board for 4chan Pass users: /vip/ - Very Important Posts
[Hide] [Show All]


🎉 Happy Birthday 4chan! 🎉


[Advertise on 4chan]


File: spurdostores 8bit.png (8 KB, 300x317)
8 KB PNG
it will be english text where something peculiar about the text says its AI
humans wont know what it is

watermark will only happen on a lengthy enough piece of text, not in a short sentence like "hello world"

https://www.theverge.com/ai-artificial-intelligence/1004880/openai-chatgpt-text-watermarks-eu-ai-act

it appears in EU region at first and comes to USA later (due to EUs strict law requirement on AI)

OpenAI says its textGrain watermarking "matched or exceeded" other approaches like Google DeepMind's SynthID for text, which is also the basis for the watermarking Anthropic announced in August 2026. Like OpenAI, Anthropic made the move to meet the requirements of the EU's AI Act; however, not everyone was happy to learn about the addition.

OpenAI also included scores from AI benchmarks showing similar performance from watermarked and unwatermarked text. But it notes that textGrain "does not guarantee reliable detection," and says text watermarks don't verify accuracy, determine who owns the text, measure how much a human contributed, or prove human authorship, it only shows it was made by AI, for EU this is enough.
>>
File: 1766237756338136.jpg (32 KB, 391x306)
32 KB JPG
>>543899596
Well that means you can filter it
>>
>>543900167
you would need to know what to look for to create a filter

>its not "this is AI"

its something subtle
>>
>>543900309
If it can be detected, it can be filtered. Micro spaces, cryptography, syntax shift, etc.
>>
>>543901061
That depends on who has the "encryption key". You could make it something incredibly arcane that no one else would ever figure out (e.g. take the tenth letter, convert it to its numerical place in the alphabet, do the same thing with the 20th letter, raise the first number to that power, prime factorize it, take the second exponent, raise it to the tenth power, then take the 4th digit of the resulting number, take that letter in the alphabet, and make that the 25th letter)
>>
>>543899596
I would pay them 50 euros a month for a extension that filters out all the AI slop jeets off the internet
>>
>>543899596
It takes about 2 seconds to remove that watermarking.
>>
>>543899596
Never know what could it be.
I am considering doing it to.
Granted my intellectual property should
Get some repect too.
Either this or I'll have to begin
Raiding datacenters for scrap metal with local niggers
>>
>>543899596
I bet it would be next to impossible to remove the "watermark" by using a local AI to paraphrase openAI-generated text.
>>
>>543901915
.....and an LLM would do that instantly on fast setting. yeah, no.
>>
>>543902608
If you didn't know what you were looking for it would require you to check an essentially infinite number of possibilities like this
>>
>>543900167
Based
>>
>>543899596

No need. "Superintelligence" is noticable enough.
>>
>>543901956
Jeets need AI because they have no intelligence
>>
>>543900309
The point of it is so that it can be detected by other ai's.
So you just get an ai that filters other ai content.
>>
>ChinkLLM, take this attached text, sanitize the characters, only leaving english alphabet letters, numbers and basic punctuation intact. Then, rewrite the entire text in a slightly different writing style, keeping all names and values intact.

gg ez
>>
>>543900309
It's text Anon. The only way you couldn't notice it is if you didn't read it
>>
>>543901915
Anon, that would still be the one word that's misspelled regardless of what the number is
>>
>>543904304
Anon, it's TEXT
>>
>>543901915
This is literally encryption 101. If you have a non-lobotomized AI, it should easily be able to crack that.
>>
>>543908527
yes but you mighnt not notice one misspelled word in huge text
>>
>>543901061
Unless it's in the syntax, a screenshot and OCR would solve that problem.
>>
>>543899596
You'll never notice because it will just say "Look at me I'm a faggot"
and OP will agree
>>
>>543899596
It will literally just take a few hours/days for people to figure it out and find ways around it. If, by some strange worst case scenario, they have invented some software blackmagic that copies the watermark whenever you try to copy paste text, then that just means you have to retype the text manually.
>>
>>543899596
>in your own words
wont do much but filter out a bunch of literal retards.



[Advertise on 4chan]

Delete Post: [File Only] Style:
[Disable Mobile View / Use Desktop Site]

[Enable Mobile View / Use Mobile Site]

All trademarks and copyrights on this page are owned by their respective parties. Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.