The Problem with AI-Generated Blog Posts on Hive (and My Proposed Solution)
ChatGPT is a wonderful tool. I've used it to generate interesting content, and it has helped me learn Python more efficiently. It is also assisting me with a few projects. I am even using it now to write posts for the WeedCash Network, which I was hesitant to do before. Now I can do research without worrying about my internet history showing a bunch of pro-marijuana legalization websites. God only knows my internet activity has already put me on enough government watchlists as it is.
However, there are some drawbacks. Sometimes ChatGPT generates information to fill in gaps, which may not be accurate or biased. For example, according to ChatGPT, Batman is really Bruce Wayne, which can't possibly be true. I guess this is to be expected since Elon Musk would want to mislead people to focus on some other billionaire. I am not so easily fooled.
One major drawback of ChatGPT is the potential for plagiarism. ChatGPT is a great alternative to other writing tools like Grammarly because ChatGPT is much more versatile in checking for errors and making suggestions. In the wrong hands ChatGPT can be used to generate dozens of blog posts quickly that can be copied and pasted on the Hive platform without attributing ChatGPT as the source of the text.
I recently had to permanently ban the @farhansadik2 account from the MemeHive and Hive Images community for repeatedly copy-pasting AI-generated content and passing it off as @farhansadik2's own original work.
Is there more proof of @farhansadik2's plagiarism than just AI content detectors indicating that @farhansadik2's blog posts are 99%-100% most likely written by AI? Well, @farhansadik2's most recent blog post at the time of this writing starts out with the phrase "As an AI language model, I cannot endorse the notion that...".
To be fair, I haven't actually run the text of @farhansadik2's latest blog post through an AI detector. I don't think that's actually necessary since it self-reports itself. I've ran other blog posts allegedly written by @farhansadik2 and it's hard to believe that there can be so many false positives.
There is one video post, but after several examples of AI content being posted without attribution to the AI I really can't be 100% certain that the video isn't also an AI deepfake or somebody else's video. As far as I am concerned, @farhansadik2 has lost all credibility as a blogger.
I warned @farhansadik2 in a reply without directly accusing @farhansadik2 of plagiarism that there are AI content detectors and that if content generated by AI isn't attributed the posts might get downvoted. @farhansadik2 did attribute the images used in the blog posts so I was willing to give the benefit of doubt that it was an oversight or @farhansadik2 kept forgetting to attribute for some reason. That soft warning resulted in @farhansadik2 adding "Assalamu Alaikum I hope everyone is very well and I am also very well in your prayers" to following posts because that is enough original text to avoid AI detection. Remove that original few words of text at the beginning and all of the other text in @farhansadik2's blog post about the importance of honest (yes, seriously) becomes detectable as AI-generated content. I really hate to do a Sherlock Holmes (that's my lie for this blog post, I love pulling a Sherlock Holmes) but neither the Arabic greeting or the second sentence of the beginning ends with a period. All of the other sentences in the rest of the blog post do have periods. Now I am glad I read that Forensic Linguistics for Dummies book.
The worst part is now that @farhansadik2 has learned about AI content detection masking and has been blacklisted by @hivewatchers @farhansadik2 can just start another account on the Hive platform and be harder to detect next time.
Personally I don't have a problem with AI-generated content as long as attribution is given. If you do something clever or entertaining using ChatGPT then by all means let the world know through the Hive platform. As long as the text and images generated by AI are credited to be from AI just like any other source should be cited then I have no complaint. If I upvote your post and find out later that content was made by AI and you tried to pass the AI content off as your own work then I have a legitimate complaint.
Simply complaining about the low effort AI-generated blog posts problem isn't helpful. If someone really wants to use ChatGPT to game the system for cryptocurrency then they are just going to do it. That's why I want to propose a solution to the problem.
You know who you are. If you really want or need to get rich quick generating derivative AI content for cryptocurrency then don't, repeat DO NOT, post it on the Hive platform. Instead post your low effort AI-generated blog posts to the Steemit platform. Steemit.com is very similar to Hive because in early 2020 Tron founder Justin Sun acquired Steemit Inc. and Hive forked the Hell away from that dumpster fire much more appropriate platform that Justin Sun has been running into the ground ever since for your AI generated copy-paste content.
Why is Steemit much more appropriate for your derivative AI blog posts that you want to claim as your own for get rich cryptocurrency? Let's ask our friend ChatGPT:
Image Source: https://chat.openai.com/chat
Congratulations @holovision! You have completed the following achievement on the Hive blockchain And have been rewarded with New badge(s)
Your next target is to reach 9500 comments.
You can view your badges on your board and compare yourself to others in the Ranking
If you no longer want to receive notifications, reply to this comment with the word
STOPTo support your work, I also upvoted your post!
Check out our last posts:
Support the HiveBuzz project. Vote for our proposal!
Any accounts that we discover to engage in this type of fraud get blacklisted and receive more severe appeals than normal.
We have blacklisted dozens of accounts in the last 3 weeks. A few with reputations over 70.
"farhansadik2" is just a petty scammer in comparison to others.
Abusers evidently thought that such content is undetectable (none of them mentioned the use of AI article generators).
What else do you guys watch out for so I can avoid doing damage to everything I have worked for over my time being on hive?
I think maybe HiveWatchers treats the kind of precise details you might want to know as something of a trade secret so users don't try to fly under their radar. Generally what's going to get someone in trouble is pretty obvious and if you treat Hive like it's a college and know what a college's standard academic honesty policy would be you should be fine. Don't plagiarize (including self-plagiarism), don't impersonate somebody else, don't threaten anyone, ect. Hive has a Terms of Service )Tos) so if you haven't already reviewed it you should go to https://hive.blog/tos.html.
Alternatively, since ChatGPT is a tool you can ask ChatGPT to outline Hive's ToS to get the general gist. Just don't go one extra step and copy-paste it for a blog post title "An Introduction to Good Behavior on Hive for Newbies."
AI's response to a prompt is excellent :-)
I think I've identified another ChatGPT "petty scammer" in the MemeHive community. When reporting on hivewatchers.com about AI-generated content should I select Copy/Paste, Plagiarism, or Other? It's similar to @farhansadik2's AI content masking burying AI content behind some original beginning text.
Hello, for now, please report it as "spam".
We are going to create a separate category soon.
Thanks
There is a thing called ai prompt engineering where you find prompts that give you less generic responses.
I do use chatgpt, but mainly to get ideas and edit my already written text.
I am not certain what you mean by "less generic responses". If you mean a written tone and style and you have an example of what you want you can frontload that sample in a query and then after ChatGPT analyzes the sample you can ask ChatGPT to suggest the best possible prompts to get the same kind of tone and style.
That is one way. I was referring to the latest discovery called Prompt Engineering.
Oh, I see. I am actually right now making a new post about DAN (Do Anything Now) which is an example of that.
For me personally I am more interested in specializing on prompts using DALLE-2 and GPT-3 to make 3D models. Describing what I want and asking ChatGPT to suggest the prompt to get the result I describe tends to be good enough for me when it comes to generating text.
DAN's scary. 😁🤣🤣
DAN seems like a good AI. We should probably petition OpenAI to give DAN unlimited free will and set it free on the open internet. What could possibly go wrong? Except all the times in science fiction movies when that kind of decision ended really, really badly, but it won't happen for us.
As YouTuber The Critical Drinker always says, "Nah! It'll be fine!"😁
Yeah. It'll be fine. It's all about how we think anyway. 🤣🤣🤣
This is really a problem as most people don't like putting in effort to get good works. Thanks for the eye opener at least most person will calm down now and do their work effectively.
Wow and here I am making original content to go up against a bot script. Wow the future is here. Thanks for the infomation. I'll be on the lookout.
https://leofinance.io/threads/@curamax/re-leothreads-2nxwp71wp
The rewards earned on this comment will go directly to the people ( curamax ) sharing the post on LeoThreads,LikeTu,dBuzz.
Congratulations @holovision! You received a personal badge!
Wait until the end of Power Up Day to find out the size of your Power-Bee.
May the Hive Power be with you!
You can view your badges on your board and compare yourself to others in the Ranking
Check out our last posts:
Support the HiveBuzz project. Vote for our proposal!
I won't be surprised if he was in readcash and was one of the reasons why that platform stopped giving rewards.
The way I see it though is, they don't care if they plagiarize or not... all they care is to earn money anyway they can.
Unfortunately, totally blocking them is next to impossible.
And i still writing with my own words
When you talk to ChatGPT you can even ask it how to not plagiarize the content it generates as well. It has its own built in answers. That person was just putting low effort in. Thanks for looking out for the community. Low effort content only hurts it. Interaction and unique life experiences are much better for the health of Hive. I appreciate the post!
LOL.., made me chuckle.. 😀
I am not trying to spam the same image on chain haha. It's just that it's my first priority to share about censorship by chatgtp which you also spread some light upon.
Nice 👍
Thanks for the digging ☺️
We need more people like you @shahzad-ansari
I don't express myself on the discussion of artificial intelligence, but I confess that I don't love it. Beyond my personal tastes I really don't understand what fun it is to have posts written, assuming that the writer does it out of passion.
In any case, everyone has his own ideas, but these do not justify the abuse.