Should AI generated publications be considered fraud on Hive?
Conversations about AI have flooded the internet in recent times and for the most part, I've avoided them due to my lack of knowledge. However, due to a recent event concerning the Hive FPL project, AI caught my attention and I learnt that it is considered fraud on chain.
I don't have a soft spot for AI, nor do I have a hardline stance but I know that AI threatens my source of livelihood. The fact that anybody could generate a prompt, send it to a bot and generate blocks of text is hardly fair to people like me who spend hours of our lives trying to bring words to life.
At the same time, I understand that AI serves a purpose and one of them is to enable people without the requisite skill to easily generate said blocks of text that capture their opinion. Using the right prompt, users can almost perfectly capture their opinions about a topic and I don't think that should be taken away from anyone.
I don't think all the AI software being created are your friends but I don't think they're the enemy too. I think they're created to live in their own realm and users that generate AI generated content should stick with each other.
One thing I know for sure is that AI is definitely not going away anytime soon but it doesn't necessarily have to be accepted everywhere. AI-generated text, images and media, as a whole, are still thriving in the grey area of acceptance in the wider society and as the days go by, with more fine-tuning to the software, I expect AI to become even more human and be able to best all the checker.
Checking Checked Checker
To get an idea about how AI works, I gave it a whirl by comparing what AI does to what humans do. I have an AI checker that I've used on some applications on hive and for the most part, they've been human text.
What I wonder is whether AI can be circumvented and how difficult it actually is. To test it out, I will use the prompt "Is using AI considered fraud on hive in one paragraph" and it turns out that Chatgpt kind of sees it the way I do in that it thinks it rewarding should be subjective and the behaviour of the account, rather than the said AI text is what should be considered fraud;
Using AI on Hive is not inherently considered fraud, as there are many legitimate use cases for AI on the platform. However, if the use of AI is intended to deceive or manipulate other users or the platform itself, then it would be considered fraudulent behaviour. For example, using AI to generate fake engagement or upvotes on posts, or to artificially inflate one's earnings, would be considered fraudulent. On the other hand, using AI for tasks such as content analysis or spam detection could be seen as a legitimate use of the technology. Ultimately, it comes down to the intentions behind the use of AI and whether it is being used in an ethical and transparent manner.
Strangely enough, the AI checker that my wife recommended for me accepted this previous text to be human-generated. I checked with another AI detector and it once again found the previous paragraph to be written by a human.
Funny enough, I could easily take this idea in the italicized paragraph above and remake it entirely in my own words. It is a pretty simple procedure that I seriously doubt would have alarm bells ringing from my publication or will be detected by any AI checker.
My understanding of AI is still very limited, I've only used it twice in my life and so, I don't have vast experience in checking the "AI-ness" of text. I will continue researching the concept to see what i can learn and I will also challenge myself to write like an AI and see whether the checkers are actually good enough to detect that it was a human that wrote it.
AI on Hive
As mentioned at the start of the post, AI is a major issue on hive that could lead to downvotes and whatnot. Steemcleaners or Hivewatchers have set up scanners that go around Hive picking up posts that are generated by AI.
Now here's the thing, people will definitely use AI on Hive but receiving rewards on the network will be a completely different issue. The fact that some accounts are against the use of AI doesn't mean that this is a decision made by the community but considering that our earning on hive is dependent on our effort, it is only fair that AI generated text feel the heat but this is not to say that I'm saying they deserve it or not, I'm saying that I understand why.
What I'm saying in essence is that I don't think AI generated publication should be considered fraud because it is fundamentally not a scam and as such, individuals that use AI should not be vilified by the community. However, rewarding or taking away the reward for said generated content is at the mercy of the Hive public and that's not peculiar to AI-generated posts.
Posted Using LeoFinance Beta
In the end it still is creativity that is the start of a great post. If ai was used to write a post, but there’s good information and the ‘writer’ came up with an interesting idea for his post, then I don’t mind ai.
Of on the other hand someone is creating tons of useless facts-posts that nobody likes to read, then this should be stopped
Yes, I like this perspective as well.
On hive where people are rewarded for proof of work, I think it is fraud.
Doesn't this mean that we will stifle the voices of those who rely on AI to share their opinion through prompts?
I believe AI generated content should be allowed but not rewarded. They shouldn't be flagged to infinity, so people can see the opinions expressed.
It’s fraud to claim that something you’ve created was in fact machine generated.
In my opinion, keeping AI generated content off Hive helps to strengthen Hive’s use case to content creators and their audience.
Interestingly Google are already promoting the use of Bard to content creators - below is an email I received just last week.
I won’t be taking up the offer to trial Bard but I can imagine a lot of people will.
Ultimately the short term rewards you may benefit from will be replaced when anyone can create similar content to you with just the click of a button. Any genuine content creator will enjoy the creative process itself and ultimately audiences will be looking for unique, personal insight which AI can’t currently offer.
Well, I do understand your point of view and yes, I don't think it's fair that people can just make content with the click of a button but considering that;
Is it right to completely dissociate hive from AI?
I feel like we can find a middle ground for AI content in the same way platforms like Leothreads found a middle ground for short form content and how Liketu enables you to "publish" one picture on chain.
Perhaps prevent AI generated posts from earning on hive but allow visibilty since the information expressed is the opinion of the person that sends the bot the prompt. This way, we won't stifle the voices of people who rely on AI to share their opinion and this also alligns with the ethos of hive.
As long as you don't produce the content it is fraud...
as long as they don't cheat by claiming that it is a "pure" from themselves (even though there is some "AI generated text" inside the post), it's still OK