Topic: How Can We Identify Someone Who Uses AI? - page 3. (Read 927 times)

Check URL & Text for AI origin with Writer
Human vs AI content detector by ContentAtScale.io
GPTZero detects ChatGPT, GPT3, GPT4, Bard, and other AI models.

If you want to detect whether or not someone is using AI to generate their posts, then use either or all of these tools to detect whether or not/how much AI was used to create the post. I'd say that there is a lot of fear around using AI to make posts in comparison to when AI was first being used to make posts, though do come back with some feedback on how you went finding them.

You'll find many tools that will let you detect AI posts, but the accuracy of most of those tools' detections is debatable. Because users who post AI-generated posts mix it up using other tools to make it look like it was written by humans. Then if they mix AI generated posts with human generated posts then it becomes difficult to detect them. So it is true that it is difficult to detect AI posts with 100% certainty.

Some people may take help from Grammarly to correct their grammar, and others who are proficient in English may never make a grammatical mistake, hence I don't think grammar perfection should lead to the conclusion of AI posts.

If a person is constantly using the AI, one can see his overall posts and get the idea that the posts lack the human touch and then we may further make use of sites that detect the AI for further confluence. On the other hand, if a person uses AI in a few of his posts or in parts of his posts, then it may be difficult to detect the AI posts.

Perfect grammar may be a sign of AI usage if the said user has a history of poor English knowledge in their previous posts. Otherwise, in general, grammar isn't an accurate way to suspect AI usage; I'm also quite proficient in the English language and also use extensions such as Grammarly to correct any mistakes I make. The truth is that you cannot suspect every single newbie that creates constructive content, but the truth is that after the introduction of AI platforms, there has been an increase in suspiciously constructive threads, which may raise red flags if it's a common occurrence.

Perfect grammar can't act as a evidence to say that someone is using AI because before this CgatGPT and all others there are many newbies came to the forum with vast knowledge about cryptos as well as with good communication which depends on their communication skills not related to knowledge.

But the wall of texts with more generic opinions can be a red flag in my opinion.

Or double check with tools like zerogpt [1], I have tried copy-pasting from ai there and it detected well, and rarely what I wrote by myself. I tried couple others, but they even detected my own written content as ai so I discarded them.

---

[1] https://www.zerogpt.com

That's actually true, and I've tested a few tools with my own posts. Some showed that there was a possibility that my post was AI-generated; others showed that it is 100% genuine text written by a human; and some others showed that it was 100% AI-generated. That's a huge deviation among the few tools that I used, some of which were completely inaccurate. I've never used AI to write a simple post, even if I wanted to. I'm not that dumb to risk having my account flagged for such a stupid reason. Thank goodness, my English is quite decent, and I'm capable of writing posts myself.

Generally, AI-written posts sound too generic and robotic, and I've also noticed the usage of bullet points is quite common. I don't believe that any older member would resort to using AI just to reach their quota; you'll eventually get caught by more experienced members who'll flag your account. It's mostly newer members who think they can outsmart others and find shortcuts to entering signature campaigns, believing that they won't get caught. Thus, it looks quite suspicious when a newbie posts a detailed thread with perfect English and no other post history.

Those tools are actually not reliable as they have failed so many times. We are yet to have an accurate AI text dictating tool. But apart from using tools, I identify AI posts especially when posted by a newbie whose writing pattern I have not studied. When the grammar is so correct and meaningful, yet coming from a newbie without merits, it's always a red flag.

Op, instead of trying to dictate AI written texts, just avoid using it yourself and others will as well avoid it and we will have a better forum with human understanding.

How can you tell if the OP is babbling, or are you condemning all OP is saying from his point of view? Sometimes earning merit isn't the main thing (and how can you tell that @LMFS was going to give him a merit, you make me laugh though 😅) in this forum is getting the issue at hand cleared, so you have to clear the OP other than making him look small or something like that and remember he's a newbie who's ready to learn and arguments will always flow when it has to do with someone who's ready to learn.
Talking about the AI, some members do abuse it, and the thing is we can't tell the intention of some of us using it, we can't be always perfect.

AI generated post can be known from many clues as you’ve mentioned but I don’t think having not to do mistakes with punctuation marks should be a top priority when suspecting whether a post is AI generated or not. Good punctuation marks also signifies an AI generated post but other things have to come in place before tagging such post as AI generated post.

Users that are not good in English use other online paraphrasing tools to strengthen their words and writing skills. This tools only help to improve your vocabulary writing abilities and their punctuations are very comprehensive with the words they help in paraphrasing for you. (It is legal if they’re your words and not trying to paraphrase someone’s idea without citing them; that will be termed plagiarism and illegal). Users who use such tools can’t be labeled an AI generated post. So good punctuation marks does not signify an AI generated post.

There are some still using AI to post but i want to believe it's common to the newbies, the forum's rules and regulations does not go against it but most of the signature campaigns kicked against such, to me using AI is as if you just plagiarised which is not good enough.


They make use of AI for many reasons personal to them, some it's because they are lazy to learn and read to understand how they can present something on their own, wone feels it's more secure way to hide and do copy and paste than direct plagiarism, some is to physh for merits, some cannot make a good composition except an artificial intelligence bot do that for them.


We cannot clean them up completely because the more we do the more new members are coming to start another, except if the forum go against it.


There's a thread on reputation where you can report any AI post, they ended up being tagged there, but they don't get banned.

There's no other solution for this behavior but to report the post since there's no rules in the forum where you are not allowed to use AI to generate post which is what they will post here in the forum. As far as I know, there are forum members who are trying their best to help us how to determine posts that are generated by AI. If you ever see one, you should report the post and either you contact a DT to tell that a forum user is using AI to post or not.

Had you provided good proof to confirm that your claims are real and not just some paltry babbling, you might have been lucky to earn yourself a merit from @Lmfs. but you are only making false claims that will not lead anywhere.
Although a few members are smart enough to use the AI, believe me, campaign managers always look closely through members posts, and if they see any AI posters, they will remove the person from the campaign. Also, there are some reputable members on the forum that can identify AI content.

When is the time where AI website like ChatGPT became popular? Just so you know, it's on November of last year that it went very popular that it broke records. Administrators here aren't just propose a rule out of nowhere with regards to AI being used here though I will say that I despise, and I hate those users who are using AI websites like ChatGPT, and the like to create more constructive post here, but right now there is no offense or sanctions to users who are using it.

I expect changes though if AI became more popular, and will be used more by users here though I don't how. A negative trust maybe in the future?

Merits, and Money

This is the primary reason why they are use AI. AI creates more constructive posts, and when some users especially high ranked members here saw that post, they have a high chance to get merits, and when they get merits, they increase their rank thus, they can join in a signature campaign where they can make money.

I guess I'll just let the Administrators of the forum address to this problem, and hopefully they will come up with a solution with this one.

You can use AI detective tools but you can smell it without tools. AI contents usually are very long and they use like bullet points, long paragraphs and wording styles are like a cocktail because they use contents from multiple sources.

https://contentdetector.ai/
https://writer.com/ai-content-detector/
https://contentatscale.ai/ai-content-detector/
https://x.writefull.com/gpt-detector
https://hivemoderation.com/ai-generated-content-detection
https://paraphrasingtool.ai/ai-content-detector/
https://writer.com/ai-content-detector/
https://copyleaks.com/ai-content-detector
https://sapling.ai/ai-content-detector

There are articles/videos from prominent influencers explaining how to detect and create undetected content using an AI copywriter. nutildah has created a thread on the Reputation board which is meant to find members who are using AI on this forum. Since the forum rules do not have anything on AI generated content. It is advised to report such content to the moderators and then they would decide. Most of them caught on that thread are new age members who might have followed those influencers. After a few years, I do not think that thread will hold any value or those detectors as AI chatbots have already started evolving themselves to write & sound like a human. I read somewhere that the upcoming version of ChatGPT would be more humane than the current version.

I know is some how difficult to get what I said. I never said in anyway that there are no good "lengthy" post like you call it, they are but there are also good and lengthy post that if properly checked the poster made used of an AI to compose that post...
Before you conclude read and get what I'm driving at, I hope you'd be my psychologist 😂😂😂 it would help a lot.

Although there are tools available to help you identify AI-generated writings, sometimes you may tell just by reading a message that it was created with this technology, especially if it was made by a novice user. when a new user who has only recently joined the forum begins posting well-written, well-composed grammar that is beyond the capabilities of new users. These posts are either created by newbies using AI or by an alt account belonging to a user with a higher rank acoount. Although they are aware that it is against the forum rules, no higher rank alt will create such an AI-generated post.

There are several tools used to identify AI text. https://sapling.ai/ai-content-detector Is one of them which i feel is efficient enough, it is being used by lovesmayfamilis to report defaulters who post AI generated texts on the forum.

Free tools most of the time online means unreliable though which means that it's not an accurate measure if you want to deter AI usage in this forum, some people use grammar checker like Grammarly to make their grammar almost inhuman. I would take those tools with a grain of salt. Pro tip though, you can ask ChatGPT if they've written those statements but that only works if they use ChatGPT for their posting. To me, detecting AI generated content is difficult because sometimes it's in the way they construct the statements which the AI can copy to make it sound like a human wrote.

Detecting AI posts is complicated because there are several detectors with inconsistent results. Some tools will report that a post is AI generated while another result will show that it was written by a human. This will pose a big challenge to the community making it difficult to come to a consensus about posts. These AI tools are designed to copy humans and every day there are advancements in this technology making it difficult to dictate. The AI technological sector is an emerging market that will mature in the future. We hope to have some trustworthy AI tools that can be precise and accurate in detecting works written by bots.

Manually finding AI posts can be very difficult because some users usually edit the results to give them a human face. At first, I kept suspecting that long posters might be using these tools but it turned out to be wrong. This is because some people like doing long posts. I also began to assume that most well-written and error-free posts might be connected to these tools, but it also turned out to be false. I have also seen members that were wrongly accused of using artificial intelligence to write posts because their post was long, looked generic, and error-free. And when the community investigated it turned out that they were written by humans. The bottom line is that there is no known standard to rate AI-generated posts.