Is @steemitblog's Post Also AI Generated?
I would first like to draw the attention of the moderators and admins of all the communities that are currently holding the Steemit Engagement Challenge.
This is an issue that has been facing by everyone since the invention of GPT. And there is an urgent need to sit together and find a solution. Several friends have had this complaint, and it happened to me twice.
So thought to inform all you moderators about this. Readers are requested to spread the post as much as possible. Remind me if I forget something.
Mention With Polite and Respect:
@steemitblog, @steemcurator1, @steemcurator01, @adeljose, @rme, @hive-181136, @fjjrg, @rmm31, @colombiaoriginal, @crypto-academy, @franyeligonzalez, @marpa, @steem-cameroon, @saxopedia, @fombae, @chant, @majerius, @fonjougiresse, @hive-126193, @steemithealth, @steemitphcurator.
If I have the audacity to mention all of you here, then obviously it must be a necessary Reason. Please take two minutes and watch this video.
Hope you'll don't mind with quality or sound quality. But surely you will try to understand the point.
There are much bigger problem than this specially have a look on these screenshots:
|
|
This is my recent post SEC s8w3 Elections in my Country .
I wrote the post with my 100% and my hard work. I got a comment by dear respected mode
@fonjougiresse that my post is 7% GPT generated. (As seen in the video, then my percentage decreased accordingly. 😂)
Wait here's another big example, iys
|
|
And this is another tool, not zerogpt. And it's showing that this content is having 11% generated material too. 😂
Last Example:
|
|
|
And now, if you check the content of this post of mine in a detector, it will still give you some percentage for AI GPT. Never give a hundred percent for human creation. 😂 And this thing creates differences۔
My job was to put up with the problem that I put up. Now it is the job of all you senior admins moderators and especially the steemit team to think of a solution.
I am going to sleep now, 😴.
All I would ask after this post is for a moderators not to tell me 7, or 8, or 20 percent of my post has been generated. Because I have shared the 65% user with you. And when my post doesn't contain such material, and you insist to me that it does, that's wrong.
Goodbye;
It's @growwithme©
Special & Original Work!
Greetings my friend @growwithme
The tools we are using to detect AI content show a probability, therefore, it is important to take as a reference the results above 70% or 80%, in addition, the same content should be applied in other tools to corroborate the results in order that our judgment is well supported.
I will try in the next hours to explain a bit my search method and answers before these cases, maybe you will be interested @steemcurator01.
In your case, 6% is not a reason to label you as an "AI User", as it is a smaller percentage that relates your content to any amount of information on the internet.
@abuse-watcher
That's what I had to say to these moderators, as you can see I also replied to them. But there was no reply, and only because of that my post was given a score of 6.5 so that I would not be able to stay in the contest.
I just want a solution that helps us all. And I would like to draw the attention of moderators to what you have done.
Thank you very much for your comment on my post.
It is completely wrong to accuse you of generating AI content in your post
6% is too low for them to compare this with AL produced
I wonder what the future will be like as we humans use many words in our posts and of course some sentences will be almost the same. Especially when it comes to professional literature.
I used a number of services in the past when I had the competitions
Some of my users who take part in my competitions are skilled at writing such stories and I know that they did not cheat, but still when I checked the posts it came up with plagiarism of 30-40%
So I would say that when the databases get big enough that store words and sentences, the % will increase so that it can appear that what we write can look like plagiarism
I am very grateful for your very detailed comment. The real thing is trust, as you said you trusted the story writers weren't cheating. Still some results come up when you check, so obviously that doesn't mean you should dismiss them as cheaters. As @adeljose has also commented above, it is a wrong move by the moderators to declare the minimum amount like this. They should consider 80, or 90%.
Greetings Friend @xpilar hear is @fonjougiresse please how do you deal with the issue because I also have the same problem with plagiarism detector tools. Even me, when I finish writing my post I run my own post on plagiarism detector tools and it still find some plagiarism percentage.
Greetings friend @growwithme. You have mentioned many people in your post including me that is why I was not notified and responding now. Another mod send me the link today and luckily i saw your post and went through it.
Friend, I am very sorry if I offended you in one way or the other with my comment in the community, I am fully responsible for that, no one else is responsible.
You are correct this tool is not reliable as it always detect AI generated content. It is only after your comment on my review that I realise it and honestly I was so confused that I did not know what to reply.
But friend, now I have an idea on how to deal with it
Not only the tool I used is not reliable but also other tools.
Friend,
I know you are just kidding but you are not a poor user, if so then I am also one and I understand your confusion, for I am also confuse, we are not left behind
Thank you for a wonderful reply buddy. There's no issue not problem at all.
I just wanted to let the issue public, everyone must know this.
First of all, it has already happened to me during self-tests that my own texts were "recognised" as AI-generated. So my trust in these free tools is, shall we say, damaged. Nevertheless, they can be a helpful tool if you understand how they work and generally grasp the dynamics of an AI like ChatbotGPT. We all use common phrases on a daily basis because they are part of the way we communicate. The AI is fed lots of texts that also contain these phrases. It "learns" them and uses them accordingly. Of course, the result can then be called AI-generated. For a tool, however, it is not possible to distinguish whether a common phrase comes from the pen of an author or from a chatbot.
There would be a safe way to actually detect AI-generated texts: in the paid PRO version of the Classifier from openai.com, signatures are read out that ALL AI-based text output machines incorporate into their texts. This is not about phrases, but algorithms (word sequences with certain numbers of letters, words without certain letters, etc.).
Whether it is worth the cost and effort to acquire such special tools, whether it adds value to the steem, I dare to doubt...
Als Erstes: mir ist es bei Selbsttests auch schon passiert, daß meine eigenen Texte als AI-generiert "erkannt" worden. Mein Vertrauen in diese kostenfreien Tools ist also, sagen wir, lädiert. Nichtsdestotrotz können sie ein hilfreiches Werkzeug sein, wenn man versteht, wie sie arbeiten und generell die Dynamik einer AI wie ChatbotGPT begreift. Wir alle verwenden am laufenden Band gebräuchliche Phrasen, weil sie Bestandteil unserer Kommunikationsweise sind. Die AI wird mit Unmengen Texten gefüttert, die also ebenfalls diese Phrasen enthalten. Sie "lernt" sie dadurch und verwendet sie entsprechend. Natürlich kann das Ergebnis dann AI-generiert genannt werden. Für ein Tool ist allerdings nicht unterscheidbar, ob eine übliche Phrase aus der Feder eines Autors kommt oder aus einem Chatbot.
Es gäbe einen sicheren Weg, tatsächlich AI-genierierte Texte zu ermitteln: in der kostenpflichtigen PRO-Version des Classifiers von openai.com werden Signaturen ausgelesen, die ALLE AI-basierten Textausgabemaschinen in ihre Texte einbauen. Hier geht es nicht um Phrasen, sondern Algoritmen (Wortfolgen mit bestimmten Buchstabenanzahlen, Wörter ohne bestimmte Buchstaben etc.)
Ob es die Kosten und den Aufwand wert ist, solche speziellen Tool zu erwerben, ob es dem Steem einen Mehrwert verleiht, wage ich zu bezweifeln...
Thanks for sure. I am happy with the reaches of my post now. My work was to point out the problem we are facing in these days.
I am amazed on our moderators, (those who uses these websites, tools to check the post, text) they really don't want to give even a read manually to the post by their selves!. So disappointing is this, than how we are a good steemit user, and how we are a blogger. If we don't read others we can't think we will be read by others.
Thank you, friend!
I'm @steem.history, who is steem witness.
Thank you for witnessvoting for me.
please click it!
(Go to https://steemit.com/~witnesses and type fbslo at the bottom of the page)
The weight is reduced because of the lack of Voting Power. If you vote for me as a witness, you can get my little vote.
Your post is manually rewarded by the
World of Xpilar Community Curation Trail
STEEM AUTO OPERATED AND MAINTAINED BY XPILAR TEAM
https://steemit.com/~witnesses vote xpilar.witness
Ridiculous! No, neither you nor your observation results... ;-)
Yes, obviously wow, I can't believe it's your reply.
Thank you for your appearance.
That irritates me now. Why should I not answer like that?
You know, I'm a long-term Steemian who primarily curates with his investment. At the same time I have the (unfortunately become very rare) ability to read everything I vote. Believe me that the development of the behavior of some "important" (taking themselves very important?) moderators, taking out the "qualification" of texts, by people who can not write a rudimentary decent English sentence themselves without a translation program, is truly a thorn in my side. If these people now also rely on machines, this is for me only the sad proof that some "curators" have lost, on the one hand, the natural understanding of language, and on the other hand, the real interest in reading texts thoroughly.
But - and for this the curators and moderators really can't do anything - the saddest thing is that one MUST become so suspicious here, because too many users take advantage of this new possibility of cheating.
Haha, I am sorry for that.
Actually I was feeling special because of your comment. And now it's the detailed comment here, and I agree.
I just wanted to advise curators and moderators to read posts manually as you do. And in my opinion it's not impossible, because you know there are lot of mods and curators.
Again thank you very much!
You have made me curious. And also an unpleasant reading experience this morning, caused me to play around with the AI checkers for a while. I am outraged!
Just keep me from writing a post about it myself, because then no eye (of many "gentlemen") would remain dry!
Don't want to bother you. But wow, what a style.
Hi @growwithme
I only had a chance to read this post now and I'm not sure what to think. Issues related to AI generating content are a new problem. One that many moderators and community leaders will have to face sooner than later.
I think that it's crutial to build awareness within communities, that using chatgpt during process of content creation as a help is okey, but posting content created mainly by AI is not welcomed and it's one quick way to lose support. It's not much different from plagirizing in my opinion.
Yours, Piotr
Thanks for dropping your valuable comment here. That's right, short and bright reply.