3 ms·

That's not how it's actually going to play out. As the AI generated content becomes dramatically overwhelming in scale, the human content will become increasin

That's not how it's actually going to play out. As the AI generated content becomes dramatically overwhelming in scale, the human content will become increasingly easy to spot (and there will be multiple cues to the human content that make it fairly obvious). There will be a crossing of the two along the way, in regards to the amount of content generated, a relatively brief time where it will be difficult to tell which is which. The more AI content there is, and the more it advances, the easier it's going to be to play spot the human. The time in which it'll be hardest to tell them apart, will be in the middle frames rather than in the later stages.

317070 3y ago
Because the human data will become "stupid" in comparison, is that the reason? I don't understand why spot-the-human will become easier otherwise.
- HideousKojima 3y ago
  The human content will contain the n-word or other shibboleths.
  - chaxor 3y ago
    This also doesn't make any sense, as it would be much cheaper to use local LLMs to generate text, making it likely to see generated toxic language.
  - jerf 3y ago
    The "n-word" or any arbitrary other shibboleths may guarantee you didn't use OpenAI services, but they don't have any sort of lock on LLM technology and it's trivial to ask non-controlled ones to use any you like.
  - sorokod 3y ago
    AI generated text is still that even with human shibboleth decorations added
  - George83728 3y ago
    Deep irony to use such a dehumanizing word as evidence of humanity.
  - EVa5I7bHFq9mnYK 3y ago
    Its easy to sprinkle a few n. or other words onto AI generated content.
tjoff 3y ago
Like Kessler though the problem isn't to detect it but to get anything through. We might not see human content for a few decades before we've sorted it out.
- AndrewKemendo 3y ago
  Exactly! I just point to the root lack of differentiation as intractable As in you can’t sort it out…just like so far there’s nobody that really knows how to clean up space long term
jerf 3y ago
"As the AI generated content becomes dramatically overwhelming in scale, the human content will become increasingly easy to spot (and there will be multiple cues to the human content that make it fairly obvious)" I feel like you're not accounting for the amount of that AI content that will be deliberately and intelligently intended to masquerade as human. Every signal you can think of and may start using is a signal that intelligent humans can and will forge. And they'll get the ones you didn't think of either, because that's their job. I don't even have to hedge or qualify this prediction, because we have decades of experience with people already forging every signal that is technically possible to forge, and pouring very substantial effort into doing so, right up to having dark businesses that provide these things as a service. "Forge all the signals that are technically possible to forge" is a product you can buy right now. For example, from 13 days ago: https://news.ycombinator.com/item?id=36151140 https://news.ycombinator.com/item?id=36151140 I don't know what signals you think you're going to see, but I'll guarantee A: yes, you will indeed see them because there's a lot of degrees of competence in the use of these systems (I still periodically get spams where the sender sent out their template rather than filling it in with a mail merge), but B: those will just be the ones you notice, not the entire population. Rather than a binary decision, I like to measure with "how much information does it take for me to detect a forgery". For instance, things like modern architecture renders can absolutely fool me as being real at standard definition, but at 4K they still struggle (too precise, too clean, even when they try to muss it up on purpose). I need a few paragraphs of GPT text to identify it conclusively, and that's just the default tone it takes. Ask it to give you some text in a specific style and I don't know how much it would take. For all I know I'm already missing things. You're probably already missing things you didn't realize.
- willcipriano 3y ago
  ChatGTP has a number of topics it will refuse to discuss so any writing on those topics will be more often than not human generated.
  - jerf 3y ago
    As I mentioned in another post, OpenAI does not have any sort of lock on LLMs. There's already plenty in the wild and they're only going to get better. Already the commercial AI companies are expressing concern about how much faster the truly open stuff is developing than their stuff. The companies can still throw better hardware at the problem somewhat more reliably but that's not much of a moat; it's still commodity hardware, it's just a bit much to drop on just being able to LLM better if you're not being paid to LLM.

317070 3y ago

Because the human data will become "stupid" in comparison, is that the reason? I don't understand why spot-the-human will become easier otherwise.

HideousKojima 3y ago
The human content will contain the n-word or other shibboleths.
- chaxor 3y ago
  This also doesn't make any sense, as it would be much cheaper to use local LLMs to generate text, making it likely to see generated toxic language.
- jerf 3y ago
  The "n-word" or any arbitrary other shibboleths may guarantee you didn't use OpenAI services, but they don't have any sort of lock on LLM technology and it's trivial to ask non-controlled ones to use any you like.
- sorokod 3y ago
  AI generated text is still that even with human shibboleth decorations added
- George83728 3y ago
  Deep irony to use such a dehumanizing word as evidence of humanity.
- EVa5I7bHFq9mnYK 3y ago
  Its easy to sprinkle a few n. or other words onto AI generated content.

tjoff 3y ago

Like Kessler though the problem isn't to detect it but to get anything through. We might not see human content for a few decades before we've sorted it out.

AndrewKemendo 3y ago
Exactly! I just point to the root lack of differentiation as intractable As in you can’t sort it out…just like so far there’s nobody that really knows how to clean up space long term

jerf 3y ago

"As the AI generated content becomes dramatically overwhelming in scale, the human content will become increasingly easy to spot (and there will be multiple cues to the human content that make it fairly obvious)" I feel like you're not accounting for the amount of that AI content that will be deliberately and intelligently intended to masquerade as human. Every signal you can think of and may start using is a signal that intelligent humans can and will forge. And they'll get the ones you didn't think of either, because that's their job. I don't even have to hedge or qualify this prediction, because we have decades of experience with people already forging every signal that is technically possible to forge, and pouring very substantial effort into doing so, right up to having dark businesses that provide these things as a service. "Forge all the signals that are technically possible to forge" is a product you can buy right now. For example, from 13 days ago: https://news.ycombinator.com/item?id=36151140 https://news.ycombinator.com/item?id=36151140 I don't know what signals you think you're going to see, but I'll guarantee A: yes, you will indeed see them because there's a lot of degrees of competence in the use of these systems (I still periodically get spams where the sender sent out their template rather than filling it in with a mail merge), but B: those will just be the ones you notice, not the entire population. Rather than a binary decision, I like to measure with "how much information does it take for me to detect a forgery". For instance, things like modern architecture renders can absolutely fool me as being real at standard definition, but at 4K they still struggle (too precise, too clean, even when they try to muss it up on purpose). I need a few paragraphs of GPT text to identify it conclusively, and that's just the default tone it takes. Ask it to give you some text in a specific style and I don't know how much it would take. For all I know I'm already missing things. You're probably already missing things you didn't realize.

willcipriano 3y ago
ChatGTP has a number of topics it will refuse to discuss so any writing on those topics will be more often than not human generated.
- jerf 3y ago
  As I mentioned in another post, OpenAI does not have any sort of lock on LLMs. There's already plenty in the wild and they're only going to get better. Already the commercial AI companies are expressing concern about how much faster the truly open stuff is developing than their stuff. The companies can still throw better hardware at the problem somewhat more reliably but that's not much of a moat; it's still commodity hardware, it's just a bit much to drop on just being able to LLM better if you're not being paid to LLM.