[New Feature] Beta Testing AI Detection for Unwanted AI Content
- Thread starter PetraAnn
- Start date
-
- Tags
- ai detection new feature
Lifeisnotgood
Power Member
- Jun 27, 2016
- 535
- 382
I noticed If I use a Google Translate and copy the text from there to the thread, I think it will detect it as AI content 
So, less Google Translate Copy&Paste, then the comment should go through..
So, less Google Translate Copy&Paste, then the comment should go through..
- Jun 9, 2020
- 6,574
- 3,265
This would certaintly help the community because those AI responses aren't useful to anybody, and they generate the same piece of content that has been around for years. So nothing to learn new things from those responses.
- May 2, 2010
- 9,937
- 16,034
Is the feature live? I havn't seen any posts that are flagged as ai, or know what to look for if they are? And is it something that if the user post is suspected as ai, they have to do a captcha and solve or something before they post?
- Aug 21, 2016
- 3,584
- 831
This step is very important and necessary in today’s scenario, and I fully support this move.
- Nov 1, 2024
- 1,182
- 4,411
The feature is live. Unfortunately, I can't give much more away right now. Just remember it's going into its 2nd day of beta testing.
theusualkeysersoze
BANNED - Failed to resolve a dispute
- Oct 11, 2025
- 901
- 510
@SettingSun77 suggested here I leave a feeback while chatting here https://www.blackhatworld.com/seo/who-even-knew-that-ai-bots-almost-kill-this-forum.1800358/page-3#post-20559455
My pain point:
" If I may, can I ask for uhm an additional feature? I am sure that everyone is equally infuriated about it, says a wtb or haf thread, somehow the AI generated replies slips through, which is fine, I understand, its a beta test, but can you maybe implement a veloticty limit? Like, once a thread goes live, 5 mins before a reply can be submitted, thats what itd take a human to read and compose a message and then some sort or random 1 to 3 minutes no submission woindows for every consequent post? that would heavily disprut the most unsophisticated bot actors and even the field out a bit."
Example case:
https://www.blackhatworld.com/seo/looking-seo-backlink-for-non-stop-casino-keyword.1801371/ check the time stamp on the OP, then check the time stamp on the replies, too fast, clearly botted and some AI hints, I see, cursive is mine:
- hands-on “non stop casino,” authority-driven my apostrophes look like = "" chatgtp iuses the one to the left, I would add that as one of the filter params
- “non stop casino.”
The focus is always on real sites with stable metrics, organic traffic, and clean histories, rather than spam networks. Links are placed manually and remain live long term.
We have Sales thread of Premium EDU and DR90 Sidewide Backlinks with 50% discount offer! = half the stuff bolded, nobody is doing that, same apostrophe details
- Our links are permanent, ********, and properly indexed in Google, with no temporary setups. I also have prior experience working with casino-related SEO campaigns, so I understand how to balance aggressive keywords with safe link-building practices. = here the bot is hallucinating worse than Alber Hofmann ( ********,
These patterns repeat fiurther down, theres a real humna replz sprinkled here and there, but zeah
My pain point:
" If I may, can I ask for uhm an additional feature? I am sure that everyone is equally infuriated about it, says a wtb or haf thread, somehow the AI generated replies slips through, which is fine, I understand, its a beta test, but can you maybe implement a veloticty limit? Like, once a thread goes live, 5 mins before a reply can be submitted, thats what itd take a human to read and compose a message and then some sort or random 1 to 3 minutes no submission woindows for every consequent post? that would heavily disprut the most unsophisticated bot actors and even the field out a bit."
Example case:
https://www.blackhatworld.com/seo/looking-seo-backlink-for-non-stop-casino-keyword.1801371/ check the time stamp on the OP, then check the time stamp on the replies, too fast, clearly botted and some AI hints, I see, cursive is mine:
- hands-on “non stop casino,” authority-driven my apostrophes look like = "" chatgtp iuses the one to the left, I would add that as one of the filter params
- “non stop casino.”
The focus is always on real sites with stable metrics, organic traffic, and clean histories, rather than spam networks. Links are placed manually and remain live long term.
We have Sales thread of Premium EDU and DR90 Sidewide Backlinks with 50% discount offer! = half the stuff bolded, nobody is doing that, same apostrophe details
- Our links are permanent, ********, and properly indexed in Google, with no temporary setups. I also have prior experience working with casino-related SEO campaigns, so I understand how to balance aggressive keywords with safe link-building practices. = here the bot is hallucinating worse than Alber Hofmann ( ********,
These patterns repeat fiurther down, theres a real humna replz sprinkled here and there, but zeah
- Apr 21, 2025
- 390
- 101
I’ll type my text in the chat box first, then use Google to translate it into English. But when I paste it into the reply box, it’s easily identified as AI-generated content. What should I do? Do we really have to write it out again in English?
- Nov 1, 2024
- 1,182
- 4,411
We're aware of what's going on in the WTB/HAF sections. It's less of an AI issue and more of a different issue. But, your suggestion is noted.
Unfortunately, I can't give you a whole lot in the way of tips here except don't write like AI. Even if you're using AI as a tool for editing and formatting, you have to make sure it's not adding things or re-arranging text in a more "ai fashion." Or most importantly, taking what you wrote and hallucinate more text which it does do. This is why it's such a bloody pain in the butt on a forum like BHW.
theusualkeysersoze
BANNED - Failed to resolve a dispute
- Oct 11, 2025
- 901
- 510
Thi
@PetraAnn
Hehe thanks, like the wording I think we know what that "different issue" is, it's composite and I trust you as marketing folks will find a way;-)
S shouldn't be happening with plain Google translate at all, sounds a bit dubious to me, besides this post I've quoted looks perfectly humanly written, seems good.
@PetraAnn
Hehe thanks, like the wording
I think we know what that "different issue" is, it's composite and I trust you as marketing folks will find a way;-)Counterboy
Newbie
- Dec 29, 2018
- 48
- 19
appreciate it, there is a lot of AI slop on this forum since ChatGPT came out.
CodiBot
BANNED - Failed to resolve a dispute
- Aug 29, 2007
- 739
- 332
my post going on moderation for AI but i am using Grammarly to correct my words - as i am not native English speaking user.
https://www.blackhatworld.com/seo/have-you-ever-taken-legal-actions-against-a-hosting-company-how-possible-is-it-to-succeed.1804848/?pending_approval=1
https://www.blackhatworld.com/seo/have-you-ever-taken-legal-actions-against-a-hosting-company-how-possible-is-it-to-succeed.1804848/?pending_approval=1
- Apr 13, 2025
- 454
- 208
I use Grammarly too, and sometimes I have the same issue. I contacted support, but there is no fix for that. The AI detection is BETA, and I hope it will be better soon.
- May 2, 2010
- 9,937
- 16,034
Why you using Grammarly though? Can't you just do your best, and type what you say in your brain, to yourself? this is an english forum. I don't use grammerly anymore. Right click and browser spell check is fine. I find it strange that people need to type their message into something then put it through a washing machine (Grammarly) and expect what they REALLY mean come across with passion, and interest. I see many of you use Grammarly you can tell by your boringness use of language concepts and metaphors., and what you talk about. Grammarly sucks. Save yourself some cash, and upgrade your hosting or something else! lol
Personally i prefer to engage with people that talk real, and don't need to regurgitate their language through 3rd party tools. (even if it is flawed slightly english. )I'm NEVER getting caught out by this Ai detector on BHW. Those that do, are not creative enough with their language that's all. Grammarly certainly WILL not make your posts more creative and interesting. It just takes what you says and gives you the most politically correct perfect way of saying it the same way as everyone else. That's its job. To give you boring cookie cutter text replacements. THATS why your getting your craparmmarly posts flagged.
Can I get a flag for my posts that say 100% biological each time I post! ? lol
- Nov 12, 2024
- 1,419
- 4,033
We're an English language forum but we're not asking that everyone's English be perfect. Far from it. I suspect 95% of you would communicate absolutely fine without Grammarly or AI.
- Mar 23, 2026
- 103
- 45
When I write serious content or comments and put my brainpower into it, I get flagged as AI-generated, simply because I use GPT chat translation. But when I write something casually using Google Translate, everything is fine. Then I get flagged as spam and warned that my account will be deleted if I violate the rules again. So what should I do next?
- Nov 1, 2024
- 1,182
- 4,411
Personally, I'd go the google translate route. We're not looking for perfection here. If someone is in the DM space, one should be able to think about and format their thoughts w/o help from a LLM.
So, why we don't mind people using AI as a tool, they should be able to still put enough of themselves into it that it doesn't detect as AI. Sometimes this has to be done at the post AI generation phase vs. blind copy/paste.
growthcraft
Newbie
- May 2, 2026
- 8
- 3
Honestly this was probably inevitable once forums started getting flooded with copy-paste AI replies. The difference between using AI as a helper and using it to mass produce low effort posts became really noticeable lately.
- Nov 14, 2011
- 1,453
- 1,623
7th or 8th time in the last calendar month. Yet members using AI for every reply does not flag.
Can we just accept the fact that this module is not fit for purpose. It's made things worse (which was a very high bar).
No matter how many times you try and bang the "work in progress" drum, it will not change the above. Get rid.