This is a cross post of @azircon/i-do-not-support-ai-generated-text-as-hive-post by @kronias.
A Warning & General Announcement For The Members Of Hive Learners Community
This is a cross post of @azircon/i-do-not-support-ai-generated-text-as-hive-post by @kronias.
A Warning & General Announcement For The Members Of Hive Learners Community
I agree that, generally speaking, a person "resorting" to the use of text-generating AI (LLM) like ChatGPT are "only after the money" and/or simply using it to generate their content for them (saving them virtually all effort from making a post), BUT there will be examples of people who use this AI solely for improving the quality of their unique ideas/ content, or more as an "editor" (to improve the grammar or "delivery" - to include things like conciseness, tone, formatting, etc).
I say this with absolute confidence, because this is PRECISELY what I'm doing on my two blogs (https://hive.blog/@soulful-ease and https://hive.blog/@ai-promptwiz), and I even go the extra mile to inform my readers in EVERY post that I use AI (both MidJourney and ChatGPT, specifically) to help improve the quality of delivery of the content (my content, that is edited by ChatGPT). In fact, @ai-promptwiz is dedicated to showing all ChatGPT prompts used to help make the content of all @soulful-ease blogposts (I call them "prompt audits"), so as to be transparent about exactly how I'm using ChatGPT to deliver the content (and MidJourney to deliver any photos used within the blogpost).
To drive home the point, I spend anywhere from three to six or seven hours to complete a blogpost on my @soulful-ease blog, from start to finish, which includes coming up with the ideas for the content (which, btw, comes from a place of DEEP passion within me - its content that I love to share and feel compelled to share, because I believe that there's REAL value that's being delivered), then planning out and executing on the "prompt-engineering" side, in order to "fine-tune" the AI to make the content match my vision for what I'm aiming at delivering, then having the AI offer me a "rough draft" (which will inevitably NOT match my vision for the blogpost), then going back and editing sections withing that blogpost that don't meet my standards (which generally involves prompt-engineering all new prompts in order to do so), and only then, finally, hitting the "send" button.
In other words, it ends up being MORE (and I mean A LOT more) work for me than simply creating the whole post myself, because, in almost ALL cases, all the ideas/ content are actually my own (I actually physically complete a blogpost), but I then have to design instructions to ChatGPT on how I want it to improve/ edit that content (requiring at least 15 minutes, but likely much, much more), and then I always end up spending at least another 30 minutes (likely much more) going back and either hand edititing or prompting ChatGPT to help me edit sections of the version that ChatGPT generates for me (based on having it edit my original content). It ends up being SEVERAL steps longer than what most bloggers go through when going through the process of making a blog post.
On top of that, talking about my @ai-promptwiz account now (which is geared more towards the side of informing people on how to optimize prompt-engineering, from the perspective of my own experiences with going through the process of trying to optimize prompts in helping me to increase the quality of my own blogposts), I take extra time to write down my thought-process and reasoning behind the development of every prompt that I use on the @soulful-ease side, to include every detail down to why I used the prompts in the order that I did (usually to help "evolve" the final prompt into something that's likely to be more optimized than if I was less organized in that order), to why I used specific words/ instructions within the prompt, to any other detail that I deem someone trying to learn prompt-engineering would glean any actual value from (to help them get better at it). So I unnecessarily go out of my way to include details that I think will be valuable to others in an area that I believe needs to be given a lot more attention, because AI is certainly here to stay and LLM like ChatGPT are only going to get stronger and be used more often by people for performing their everyday tasks, so we all stand to benefit from learning how to properly prompt-engineer (which is my main drive for providing that type of content, not so much the "prompt auditing" that this account is doing for my other account).
Makes sense. I support you.
Similar to text content, we should also look at images being used. I myself has lost my photographs to AI and against licenses ...
I won't use ai generated text for the simple reason it is flagged by grammarly as plagiarism, at least chat gpt is and I have worked too hard on my reputation to wreck it by downvotes.
And since you are turning down votes on this post all I can do is toss you some !PIZZA.
I gifted $PIZZA slices here:
@wanderingmoon(1/10) tipped @kronias (x1)
Join us in Discord!
Can you show an example of some AI generated posts?
You can find examples here https://peakd.com/hive-167922/@bobinson/chatgpt-and-hive-blockchain-discussion
I feel it's so obvious when someone is using AI, U can just feel the lack of energy from people it feels like nothing but facts and information no love!!
At least this part made me happy, because while I use AI-generated content, I usually try to be creative about it and mention that I posted the content with AI's help. (It also depends on which topic I'm talking in, and most of my topics are too personal to be generated by AI anyway.)
Hola,comence a leer tu post y lo primero que me pregunte fue, y como lo saben???... que bueno que tengan 2 súper poderosas herramientas, el contenido creativo humano debe perseverar...
I think, people who uses chatGPT to generate text or edit it should clearly state that it was used in they post, but it is unclear for me, if they need permission from creators to post these text, because we monetize our content. You can find more in for here: https://openai.com/policies/terms-of-use
I did try to edit text and thought it would be a time-saver (even wrote a post about it), but it was adding new information, choice of words always sounded bizarre and repetitive. I gave up after some time trying to give it some specific instructions (I still was not happy with the result) and I find it more easy and faster to write texts myself. So what if my English is not as good as native speakers, at least can learn from my mistakes. I still think its amazing what you can do with it, but as a I mentioned writers should let readers know if chatGPT was used in they content (and then we can decide to vote for it or not).
It must be fate that I'm seeing this for the first time.
AI-generated images. Text is easy. I have to defend what I write at work or school all the time. At school, references and citations are required for just a passing grade. At work, people have ended up working for free adjusting their projects/presentation to align with company policy.
There's no reason I can't plag-check the images I'm using on AI-generated images. I've done it plenty of times for other people's work. I can also add a comment stating what I've used in my checks. At this time, I feel I've met the bare minimum for AI-generated images: Name of the image I published, the image generator, and my account name. I must do more.
Really good points here. I'll add the plagiarism checks to my posts'for AI-images from here on out. I already reference the free-to-use stuff from places like Pixabay. No reason I can't make things a bit easier for someone to double-check at least.
I agree with your opinion and I do not support AI-generated content or AI-generated text dominating Hive, because it is very hard for people who spend so much time just to create a post and the post made by some Hive users with the support of AI will be curated.
There's also a huge responsibility for curators to check the original content from those created through Chat GPT or any other platform.
Very correct. AI generated content is not acceptable
I agree with you sir. ChatGPT maybe useful in another way. But not in this platform where everyone aims to explore different personal thoughts and ideas.
I agree. I can say the same: I don't support AI generated content Why? Cause it isn't original contento. I think it's important to say what do we think about it.
Many excellent reports AI is not our goal. I have planned with my own skills. And here we definitely request everyone to refrain from these activities. But I always try to practice my skills and present them among the flowers. I love this lively letter of yours. Which further cites our expertise. I hope you are well and have a nice healthy day with your family. Thank you very much for the excellent instructions. I know your contribution here is immense. Respecting all your skills and creativity
Yeaaa all well said.
Hive is a means of learning. It beyond the money.
Thanks for this write up
Maybe you should change the text. I often use https://www.typingservice.org/ where you can get expert typing help. And you will get quality content.
It is a good measure, which should be applied to everyone without exceptions. I have seen some fairly large accounts that use AI to generate their texts, almost 70%. I simply decide not to vote for them, it is dishonest, especially on a website where a good part of the population generates original texts taken from their minds.
People actually do that! Not a slightest beat of of sincerity.
After all these months not for years, being here, people still do not find writing your own thoughts and stories are fun and worth taking some time to put your thoughts into words!?
Guess I would never understand that!
I have learned a lot from the comments of almost everyone in the comments section. this is already serving as a guide for me who is just starting. indeed it is good to read to learn from other people's work on this platform. thanks, everyone for stating your opinion.
I agree with you on this, this is a place where we learned and also share knowledge about real life content but creating a content with computer doesn't feel real. And using AI generated content cannot help you develop yourself in writing. When you write yourself , you will be able to think more deeper and come up with good content.
Creo que las reglas en Hive son claras, me cuesta entender porque se busca el camino fácil deshonesto...es cierto que hay muchos elementos que pueden hacer la vida más ligera pero para hacer contenido no aplica la IA, simple!
I agree!
Well said 👍