264
OpenAI has built a text watermarking method to detect chatgpt written content
(www.tomshardware.com)
this post was submitted on 11 Aug 2024
264 points (95.5% liked)
Technology
58111 readers
4076 users here now
This is a most excellent place for technology news and articles.
Our Rules
- Follow the lemmy.world rules.
- Only tech related content.
- Be excellent to each another!
- Mod approved content bots can post up to 10 articles per day.
- Threads asking for personal tech support may be deleted.
- Politics threads may be removed.
- No memes allowed as posts, OK to post as comments.
- Only approved bots from the list below, to ask if your bot can be added please contact us.
- Check for duplicates before posting, duplicates may be removed
Approved Bots
founded 1 year ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
The arstechnica article speculated it was more of a pattern of words thing.
I think it is lies, and doesn't exist or work anywhere near as good as they claim. Or, its incredibly easy to bypass.
https://arstechnica.com/ai/2024/08/openai-has-the-tech-to-watermark-chatgpt-text-it-just-wont-release-it/
Research on this topic exists, and it is possible to alter the output of an LLM in minor ways, that statistically "watermark" the results without drastically changing the quality of the output. OpenAI has probably implemented this into ChatGPT.
https://www.youtube.com/watch?v=2Kx9jbSMZqA
I think the tool exists, and is (at least close to) as good as they claim it is. They can't release it, because once the public can tell with high accuracy whether ChatGPT wrote some text, another AI can be developed to circumvent detection from this method, making the tool useless.
That is a long video, is the paper published somewhere?
Im willing to accept that you can statistically "watermark" the text, but I'm not convinced that it would be tamper resistant, which is a large part of what makes a watermark useful. If it can't survive an idiot with a thesaurus, its probably not gonna be terribly useful.
It can likely also be defeated by adding "In the style of X" to a prompt, changing the distribution and pattern of the responses.
...but that output is also from the AI so it would still be watermarked lol
You could feed it through a different, smaller model that could even be self-hosted. It isn't difficult to make a model that rephrases an input in another style.
Ah, okay. That's fair. It wasn't clear they meant a different system lol.
I think it exists and works but that its simply not in their best interest to have people use it and be found out that they used chatgpt, for OpenAI's business/profit potential. I have nothing to back it up but have just lost all faith in OpenAI.
Im willing to believe it exists, but not that its any good. 99% is a crazy accuracy claim.
I van totally believe that it detects AI generated content 99% of the time, that’s trivial. What I really wanna know is the false positive rate. If I write a program that flags everything, it’d have a 100% hit rate. It’d also however have a crazy high false positive rate.
Yup, noticable that they use the phrase "99.9% effective". Effective doesnt have a defined meaning in this contect, unlike accuracy, sensitivity or specificity, so that smells of missleading PR speak to me.
Especially if the claim that it's undetectable by humans.
They sell the cheating tool and the detection software
It's really sweet to play the arms race with oneself.
Not to mention that it would be extremely difficult to implement an effective watermark on text below a certain size
There are hundreds of thousands of pixels in an image where you can hide a watermark, but in a text output of a paragraph or less there are only a couple hundred characters.
How precise is the watermark? Is it a specific sequence of characters? Is it a sequence of words? A number of characters in a row? Non-print characters?
How precise the watermark is will determine how easy it is to get around. I imagine some of the most important uses to detect would be twitter/social media influence bots where the output length is only 140 characters or less. I find it hard to imagine a watermark on output of that size being effective or reliable.