- Sep 25, 2013
- 4,705
- 2,701
Google COULD detect it if they wanted (to a certain extent), however, they do not do it yet and it's very unlikely that they ever will
Google can Detect AI Content - Check this out
- Thread starter homeriscool
- Start date
A better one is this; https://originality.ai/
noellarkin
Senior Member
- Mar 14, 2021
- 1,006
- 1,492
Is AI content detectable? Yep, unedited AI content is detectable.
Can Google detect it? Probably.
Do they choose to do anything about it? Given that sites from .it network and lorem ipsum rhinoplasty sites are still around, I would speculate that no, Google won't worry too much about AI content.
Can Google detect it? Probably.
Do they choose to do anything about it? Given that sites from .it network and lorem ipsum rhinoplasty sites are still around, I would speculate that no, Google won't worry too much about AI content.
- Jun 24, 2013
- 2,404
- 3,652
What If Google is turning a blind eye to AI Content (for now) then wipes out half of the internet at a later date?
That would be like a digital nuke....
Surfdbreak
Regular Member
- Dec 14, 2021
- 300
- 266
I think I’m missing something.
Whenever I use ai, I find that it’s not usable in its raw state. It’s written poorly, and usually has a bunch of errors in it in my niche 9/10 times.
it has value to me in that I switch from creation mode to edit mode. So, I can crank out 3x the volume as before, but it’s not turnkey.
I wish I could simply drop in keywords, title and length and get finished product. Would make my life easier.
and, if I’m heavily editing everything, it would be hard to believe that Google would view it as ai.
But again maybe I’m using it wrong
Whenever I use ai, I find that it’s not usable in its raw state. It’s written poorly, and usually has a bunch of errors in it in my niche 9/10 times.
it has value to me in that I switch from creation mode to edit mode. So, I can crank out 3x the volume as before, but it’s not turnkey.
I wish I could simply drop in keywords, title and length and get finished product. Would make my life easier.
and, if I’m heavily editing everything, it would be hard to believe that Google would view it as ai.
But again maybe I’m using it wrong
Mountainsforever
Junior Member
- Dec 30, 2022
- 107
- 84
It is evidently very easy.... On another thread someone linked a spot where you could copy and paste a page and it would tell you what are the chances it was created by a human.
I ran my chat GPT pages through and it showed that they are 99.9% fake :/
But I do like the thought of Google not wanting to spend a lot more processing power and not really caring if it's fake....
Obviously that's what I'm hoping for since I have a fully GPT project going now
Mountainsforever
Junior Member
- Dec 30, 2022
- 107
- 84
Wouldn't be the first time
If my new project ranks I'm going in with the expectation of a churn and burn that at some point no matter how big and authoritative it gets Google will in fact nuke it
- Dec 1, 2011
- 3,025
- 1,921
Same, however I also took content from my blog that was written by someone who works for me and it said it was 95% fake as well. I don't think that thing is very accurate or at least not to be trusted 100%. Also, I didn't know detectors like that even existed but I saved that link so when ChatGPT API is available, you simply run what it outputs through tools like that and keep re-processing articles until it has an undetectable level. Problem solved.
- Jun 24, 2013
- 2,404
- 3,652
That's the best way to look at it.....if it doesn't happen.....great!
- May 8, 2012
- 4,218
- 3,729
as already said, the computing power needed for that is way too much. its one thing to check 1 text if its written by AI, or billions of pages daily.
they could do it sure, but not reliably enough and it would cost them a lot of money.
why would they want to anyways? as long as it provides what the user is searching for and isnt deceptive, it doesnt go against the ToS.
however at some point they will probably just announce that you shouldnt do it, as that already is enough so the whitehats and noobs wont even try.
fear tactics has always worked for google to fight spam, just scare the people who blindly belive all that google says and you already reduce lot of spam lol
they could do it sure, but not reliably enough and it would cost them a lot of money.
why would they want to anyways? as long as it provides what the user is searching for and isnt deceptive, it doesnt go against the ToS.
however at some point they will probably just announce that you shouldnt do it, as that already is enough so the whitehats and noobs wont even try.
fear tactics has always worked for google to fight spam, just scare the people who blindly belive all that google says and you already reduce lot of spam lol
- Sep 25, 2013
- 4,705
- 2,701
Some things people aren't considering.
Google COULD hypothetically speaking, figure out if something has been written by an AI or not through footprints, however, it runs into a few issues:
- There's multiple AI's out there, it's not just GPT-3 and ChatGPT despite those being the most prominent & most used. Each AI leaves its own footprint and they're all different.
- AI is meant to be constantly learning/improving. Odds are the footprints will be different some time from now, so it's a constant struggle for google to keep up. This costs time and more importantly, money.
- Computing power required to detect if billions of pages on a daily basis were written by AI or not is insane. Google reducing crawl rates for the past 2 years to cut costs already, you think they gonna invest massive computing power to go on a wild goose chase that will never end? Don't think so.
- The false positives are EXTREMELY HIGH. No matter what "AI Detection" tool you try, you will always get false positives if you feed it human-written texts. The repercussions of implementing this kind of stuff will be MASSIVE. Even authority sites will not be safe. You can never be 100% sure if it actually was AI written or not.
Last edited:
arianataylorlife
Newbie
- Feb 25, 2022
- 48
- 47
I think the main purpose of "Google" is to provide valuable and correct information to the user for their query and it doesn't matter whether the output is from the manually written content or AI-generated content.
- Oct 24, 2011
- 3,225
- 3,631
Stack Overflow can tell. They were able to detect it by around the 10th of December, 10 days after the release.
- Apr 15, 2021
- 3,120
- 1,654
It doesn't matter if Google detects it or not. Making it easier to create content will only saturate what was already saturated. It will saturate it into oblivion. The future is quality, not mass, regurgitated information. If it's easy to do, most likely it isn't worth doing it. If it's at the distance of a click, everyone will do it, and doing it will have no value. That means to succeed more will be asked from you. You will have to be better. You will have to be more creative. You will have to push the envelope.
ChatGPT won't make you rich or make people any significant amount of money in the way people around here wanna make money. ChatGPT is a redundacy machine. It will make a lot of stuff redundant and accelerate the extinction of quick money making schemes, because there wil be sooooooooooooooooooo many people doing it that it's no longer viable in any way.
Until AI can simply replace you entirely, the future is talented people offering great quality products, not kids trying to game the system with automatic generated content.
- Apr 15, 2021
- 3,120
- 1,654
It's not so good at all. It's actually quite bad and incredibly unreliable. Makes a lot of mistakes, regurgitates a lot of incorrect information, and just yesterday was giving me code that plain and simply does not exist. It will absolutely be possible to detect it, and stackoverflow alwady can to a certain extent. And even if Google can't detect it, what you create through it will just become irrelevant because the world has no use for the amounts of regurgitated information that are about to be created with this tool.
- Nov 16, 2011
- 19,799
- 11,289
They have the technology, best in the world. So they can develop something for it. Also see it like if they have to remain relevant, they will have to find a solution for this.