You are viewing a single comment's thread from:

RE: How to Detect AI Comments

I started in January to be more active. It was about 2 months ago I heard about AI-written content and the possibility to generate pictures.
24 days ago I wrote this:

https://steemit.com/hive-161155/@wakeupkitty.pal/beware-of-the-copycat-commenter

By that time I already noticed for several weeks the 'weird' comments I received followed by comments which are a copy of my own text!

I know exactly what I wrote, I recognize my writing style plus these copied comments start with an average of 5 words in badly written English. I also recognize these comments given to others because I read them.

I agree with you those AI checkers are far from 90% accurate. I tried about 5-6 on texts I knew were AI generated. Those checkers said no, or at most 50-50 but it works the other way round as well.

If it comes to generated texts I noticed the same words are frequently used plus the same rhymes and the comments lack a certain personal feeling/touch. Although, English is not my mother tongue I can recognize it by now.

I asked commenters more than once what was meant by a comment (no response), also why I am commented with my own words. I can't say I feel flattered and rather receive a thumb up instead. By now I am scolded because I dare to ask, wonder but questions are never answered.

Just like anyone who needs to translate, my translated text will have mistakes. I try to correct them if possible and noticed a translation is better if sentences are short and no expressions/idoms are used (which is hard for me).
Nevertheless it is possible to write, communicate, engage, with translators and this is what Steemit should be about.

What counts for AI-generated text should count for comments as well.
I do care less about pictures since I am a writer but thanks to @soulfuldreamer @italygame @pousinha and a worldbuilding group I am more active with it (and study the strange outcome and brainwashing hits).

I don't think 50 comments are needed to figure out if a comment is AI/plagiarism since they are different from the text/posts of these writers. But even if only two comment out of ten is generated by AI or plagiarism it is still wrong (unless it is a contest to try out something) and I am sure the commenter knows.

Since I am a mobile phone user only it would take me way more time to copy a text, paste it into another site and wait for the result, copy, switch screens/sites again (and wait till it loads) instead of writing my own comment.
If it comes to commenting I rather do it my way.
I also wrote once about how I feel about engagement contests where the winner is the one who receives most upvotes! This means I put effort in helping someone else to win no matter what is written while frequently comments are not even answered. I can't take writing creative content contests like these serious. If you want people to engage let them at least read, comment and upvote 3-5 fellow writers and let them leave those names underneath the link of their entry. Perhaps it grows into a habit to read, comment and upvote a few people per day but my experience (2018) is that will not happen.

https://steemit.com/hive-161155/@wakeupkitty/almost-too-much-proof-be-the-first-and-scribble-post-leave

All the best to everyone and if it comes to new Steemians doing this I can only say: Someone thought them that trick!

❤️🍀

#comment @ibesso @yaladeeds @el-nailul @fadthalil

Sort:  

Hello, as always, you are our light in the darkness, and your contribution always makes us reflect on what is right or wrong. I like it.

These reflections are very interesting and touch on a delicate topic that saw 2024 christen another devilry to know and understand. In a few months, it has changed the way we interact in almost all sectors.

I believe that soon AI will be integrated into everything we do, and perhaps we will no longer be able to distinguish who wrote what. Once again we will find ourselves facing a competitive scenario between generations that will have contrasting opinions about this new means of communication.

Soon our children will be educated by AI: is this good or bad? Who knows. I certainly am not in a position to judge, but it is definitely a fact.

The desire to improve performance on Steemit may push some to simplify comments to save time and comment more, but I am not sure to what extent this can benefit user interaction. For me, Steemit is a place to create connections and friendly relationships between different cultures; I don't see the profit, at least for us Westerners. This is the main attraction of the platform for me, this more human approach cannot be extended to many users because it requires a lot of dedication.

The aspect of language also plays an important role. Unfortunately, I perfected my English in London pubs and I'm not very proficient, so I, like many others, find myself having to translate concepts that often do not find the desired meaning. But this can also be fascinating; perhaps a saying from one culture, translated into English and read by a user from another culture, means nothing 😂, but this is the beauty of us humans and AI currently cannot do this.

Speaking with my loved ones, an interesting thought emerged: isn't it that all this content we write in contests is given to AI to train it to be more human, learning our way of expressing, thinking, loving, dreaming, worrying, and all our thoughts and fears?

Fundamentally, I am a progressive positivist: I believe that if AI is employed according to positive thinking, one day it could help us make humans more aware and respectful of each other, but perhaps all this is just a man-on-the-moon dream.👋

It is for now and sorry to say so but I do not appreciate being commented with my own text.
Do you use AI to generate this reply of yours?

I don't think the AI ​​knows that much about me!🤔

Unless you give it some words to use...perhaps I try it out. Thanks for the inspiration. 👋

Basically, it is easy to suspect IA-generated comments, because they all look the same way, only repeating the context of the article itself.

When you copy the comment text and paste it into the IA detector engine it will detect it. sometimes I just ignore such comments because I have no interest in reading more AI comments.

I use these three engines to check IA:

https://chat.openai.com/

https://plagiarismdetector.net/

https://zerogpt.net/id

For me; using IA and the translation engine are two different topics. Using AI is about generating words by words and constructing them into paragraphs or articles, whereas translating is actually swapping one language to another after the writer completes the article based on their thought

As I am not a native, I need to translate some words or phrases and then edit their grammar where necessary. I also have a grammar check engine just to ensure my typing, punctuation, and set of rules are correctly written, it is to avoid misunderstanding for the reader.

I think we all agree that translating isn't the issue.

We can say we all ignore the AI comments but these commenters are rewarded and grow fast while not being original.

Their posts are rewarded too since they join each engagement contest and are called heroes. Not the content is rewarded or engaging.

Just ignoring is not enough.

As you can read in the post the AI checkers are not able to find each AI generated post and they can make mistakes as well.

🍀❤️

Dear friend, in the case of texts, I noticed AI publish awful syntax and have no idea about the correct lexicon in context❤️

That is what I noticed too but it also counts for comments. By now I noticed 3 different versions.

3 different AI awful syntax versions? Yes, it's said AI will substitute humans, but what a messy job they do. Translations become garbage, out of human translators.

Yes. I gound them and can recognize them by now. Can be it depends on the AI used or the questiin adked? Still I wonder who told those Steemians to use this technique I am sure someone did. It feels as if most of them are from the same country/countries.

Can be AI will substitute humans but not in tje next decennia or two. They are slow learners,vvery slow, and tend to discriminate plus are programmed with tunnelvision. How will that work out?

My colleagues of the PTP forums, especially those from Africa, are worried about the AI substituting humans. It's understandable: they live in the poorest countries and they cannot lose their employments cause AIs fault.

There will be plenty of jobs left to check the bots since the population will be reduced and that minimum income in CBDC will be a fact to controll the people.
Here they talk about early retirement agai. Plus the war..fighting for?
My friend's daughter had a call, she told me yesterday besides we are running out of energy. AI needs it.

After seeing AI-DA, the android working as a painter, opening its own atelier, the worries increased. Robots are stealing us even artistic jobs😂. Well, in any way, after seeing the flaws branded Crayon and Freepik, maybe there's still hope😂

There are plenty of flaws. They will replace doctors firs, specialists included and the teachers.

I asked commenters more than once what was meant by a comment (no response)

This happened to me as well. Additionally, some AI commenters are unaware of what they are posting. Here is an example:

Screenshot 2024-04-27 234020.png

Cc: @alejos7ven

It’s happening for me as well, will keep trying to teach them or banning from make noise if there is not response

Will that help? Is there a blacklist. There are plenty with a higher rank/level.

How to fight this. They answer within a few minutes. I noticed it after posting. It's like talking to a machine and they rarely respond. This one is unique.

I know Steemit had a blacklist

| How to fight this.

Complicated...

I will think of another comment and will add something like "you are on my blacklist" or so

😆😆😆😆😆Ai.is generating everything now, when will it start generating food or reproducing money🤣, well if the trick makes them feel better, let them sleep with it. Originality is a personal procession that you can comfortably flaunt always. I always prefer to take breaks, than to take complete directives from AI, if someone else does the work and you are rewarded for it, how good will that be, will you sleep well knowing that it wasn't what's in your head that they found exciting. I think it was clearly spelt out on Steemit against plagiarism and I see Ai generated text as plagiarism too. That's my own take though.

The tricks help them to achieve more Steem and plagiarism is forbidden.

Funny way to live, we will all be alright 😁

Coin Marketplace

STEEM 0.18
TRX 0.13
JST 0.029
BTC 57912.75
ETH 3070.69
USDT 1.00
SBD 2.33