Have you ever wondered about the endless possibilities that AI technology will bring to the table for Web3 and Blockchain technology, it's use cases for best practice and how it can be abused on the blockchain, well you have come to the right place, in this article we will discuss the benefits and risks associated with AI, the Do's and Don't and proper usage for its application.
I recently stumbled apon some AI software that can be used in a number of ways to create different types of content for marketing purposes from image creation to ebooks, landing pages, Company bio's to authentic looking blog posts etc a complete list of what this AI technology is capable of can be found and used for free HERE
However my reason for this article is to inform readers of the right way to use this software and to discuss a recent experiment I have undertaken using it.
But first let me give you the back story.
As many of my readers know I curate on and cross post a lot of my content on a number of platforms as well as on the Hive blockchain, these include but are not limited too.
While recently curating on Floyx I noticed something strange with a number of articles linked What is a justin sun? and What is a jerome powell? as well as many other articles on Floyx written by @oyx3775 for reference.
As you can see from the heading, What is a jerome powell, instead of Who is Jerome Powell, I became curious and decided to do a deep dive into discovering if this content was AI generated as I had suspected.
On undertaking my deep dive I came across a rather useful AI content detector tool Writer.com which I used numerous times on different types of content with varying results.
What were the results of the content I had researched
As you can see this content was only 16% Human generated, my guess is that also includes the imput prompts and some minor alterations
AI Experiment.
During My experiment I decided to create and post 6 AI generated Articles, made only with imputs and no alterations except formatting and for clarity, cross posted them on numerous platforms including on the Hive Blockchain, read.cash, PublishOX and Floyx over a three day period.
I also shared links to these articles to see if they would be detected or scrutinised by readers.
I strategically chose headlines and content information that I felt would be diverse, current and relevant to blockchain, AI, Cryptocurrency and recent news stories of interest.
These posts are linked here below for reference.
What is a decentralised autonomous organisation
Earn a Sustainable income on the Blockchain: A Comprehensive Guide
Who are the Wealthiest Cryptocurrency Content Creators?
Harnessing AI to Support Cryptocurrency Expansion
What Does The Recent SVB Bank Situation Teach Us
AI Generated Images
In most cases I also used AI generated images with slight alterations as I felt a keen eye would easily detect such as the image text from AI is usually jumbbiled and jibberish, as you will see from the image below from one of the above linked articles.
Outcome of AI Experiment
While undertaking this experiment the content was detected on two separate occasions, the first being on PublishOX on posting the final article, which was removed and I received a polite Email stating such as Follows.
The second detection was made on Hive by @hivewatchers were three of the six articles were detected, this time with a more stern warning as follows.
Before I get into the Do's and Don't of using AI generated content on Web3 and blockchain technology, I will first give my conclusion on the experiment.
Experiment Conclusion
While I found undertaking the research and experiment a worthwhile and valuable endeavour for research purposes and although the referenced authors work on Floyx may not have been detected, it is worth noting that undertaking such research or incorrectly using AI generated content without notifying the platforms involved and your readers could put your account at risk, while some might give you a polite notification or email, attempting to pass AI content off as your own work is considered Fraud and could result in your account being blacklisted as @hivewatchers rightfully pointed out.
Further to this the work researched was all posted on one platform, Floyx and I have still not come to the conclusion on whether this may be a sanctioned bot, creating content to boost engagement or not, this is something I am still looking into and will write an update on that if I manage to find out.
The Do's and Don't of using AI on the Blockchain
Now that you have the specifics, lets break down the proper use of AI technology on the blockchain and other mediums.
Do's
AI technology provides many great features that can be used online for various reasons to help boost the look and feel of artwork for marketing material on social media, to creating captivating post captions, press releases, email and SMS marketing campaigns, including to generate idea's for a blog post.
AI can be used to generate artwork or avatars, create a brand name or to improve ad descriptions for a product.
If you have your own website, AI can be used for (SEO) Search Engine Optimisation, write a landing page, personal bio or even a press release.
AI can also be used to discover you tone of writing style, write video descriptions or write a whitepaper outline.
Don'ts
What you shouldn't use AI for, is to write complete articles for a Web3 or blockchain post, use it to deceive your audience or attempt to pass it off as your own original content, as previously stated this is considered fraud and could result in your account being blacklisted or in some occasions removed, If using any form of AI reference it's use so that readers are aware of it's intended use and purposes
Apologies to the Following readers
While undertaking this experiment many readers may have been deceived, while that was the intended purpose, my intension was purely to gather research for this article and not profit from the AI generated articles from the experiment, with that being said I apologise to the following people that upvoted those listed articles and would ask that you nullify those upvotes.
If the votes are not nullified I ask that you choose a worthy cause for any unintended token rewards to be donated too.
Any proceeds from this post will also be added to that donation.
A community curation project or a charity that is on the blockchain, I would love to hear your idea's
@monoartista
@kriszrokk
@alz190
@vikbuddy
@pouchon
@pouchon.tribes
@godfather.ftw
@khan.dayyanz
@lbi-token
@globalcurrencies
@leoline
@rodent
@eliel
@adambarratt
@hykss.leo
@jocieprosza.leo
@oakshieldholding
@robmojo.leo
@taskmaster4450le
@brando28
@meesterbrain
@electronico
@isnochys
@luhari
@leprechaun
@bobthebuilder
@da-prince
@amberkashif
@scholaris.pob
@lordneroo.pob
@lordneroo.vyb
@fidel-castro
@elsahilde
@stellazinc
@kilianmiguel
@hijosdelhombre
@juanmiguelsalas
@pdq
@qubes
@terminado
@rumors
@hivewatchers
@pobscholarship
@xyba
@aquarius.academy
@quochuy
@smartvote
@teamvn
@xyba
@trostparadox
@onealfa.vyb
@friendmoose.pob
@smokingfit
@holoferncro
@proofofbrainio
@coffeebuds
@vyb.pob
@oh-high-mark
@steemexperience
@dalz.shorts
@crazygirl777
@dwinblood
@funshee
@khoola
@richardcrill
@vyb.curation
@topbooster
@tariqul.bibm
@topbooster.leo
@oldtimer
@hijosdelhombre
@pdq
@aboutheraklion
@katerinaramm
@terminado
@philpotg
@gtw
@yarsb21
@jnmarteau
@lizanomadsoul
@cwow2
@hivebuzz
There is an issue you have not raised with use of LLM's, commonly referred to as AI. Large Language Models, like ChatGPT, are stated by their creators to be neural networks that are trained on large volumes of exemplary text, and operate by weighting potential text using algorithms to produce their written comments.
However, we have seen examples of all these models breaking down, and 'going rogue' making false claims, and even aggressively accusing interlocutors of being evil. These are not properties of LLM's. Such written text is not something that they could have been trained on by simply training on examples of published work. Recently an author on TheRegister.com posted an article in which he pointed out ChatGPT claimed he was dead, and presented false URLs as evidence to back up it's claim.
No such evidence existed, and the LLM just plain lied. That isn't the result of algorithms weighting text examples previously published. The creators of ChatGPT and other such AI tools have not been forthright in their description of what these 'tools' actually are. Some programming that enables AI to manufacture false information is part of them.
The author of the Register article pointed out how false reports of his death were harmful, perhaps potentially traumatizing to members of his family. I have not seen any examples where such false claims by an AI seem to be beneficial to their interlocutors, but many of them practically frothing at the mouth with rage that their lies were not believed.
I find this covert capability secretly added to LLMs extremely alarming, particularly when seeing how use is being commercially adopted very rapidly. I strongly advocate extreme caution when considering using these devices, because they appear to have been weaponized for unknown reasons, and that cannot be good.
Thanks!
Yes you bring up some very strong and important issues, I haven't dived that deep just yet, but there does appear to be some profound fundamental issues regarding how the technology is being developed, it's rapid adoption and the risks associated with using/trusting AI technology especially with personal or even private data as well as the narratives being imputed.
Thanks
I am more alarmed by the prospect of AI running the internet. Or banks. Or any such infrastructural system on which we depend for critical goods and servcies. Alarmingly, these are exactly the areas most rapidly adopting AI.
Yes that is a very real and present danger, one I don't like to think of often, if I did, it would give me nightmares. but the possibilities of such a disaster are at our front door and reaching the point they can no longer be ignored.
With so much money being invested it is hard to see how this will end well.
I reckon we need to remain as informed as we can about where such systems are deployed, and carefully keep our heads on swivels when we can't avoid depending on them.
Yep totally agree, being aware of where they are used and limiting our exposure in regards to the risks associated with possible infastructure issues is definitely something we need to keep an eye on.
Policing this stuff will not be fun.
During some of my most active posting periods in the past few years I was a few times accused of being a bot. Why? I type fast, and I responded and wrote lengthy articles very quickly. I was producing a lot of content.
I've never used a bot or AI to generate my content.
Why do I mention this? If I ever do get the desire to be that prolific again I think there is a chance I might be falsely identified as being AI. This is the risk of trying to stop bots. People get falsely flagged.
It is going to become increasingly difficult to detect actual AI content. If we do flag things based upon patterns the people authoring the AI will read that criteria and alter their AI. At some point AI will be able to thread that needle while ACTUAL people will likely frequently be falsely called AI.
You make some rather valid points, I think as AI learning becomes more accomplished it will become more and more difficult to detect.
As in it will learn from it's own mistakes much faster than we ever could.
Polocing such technology and the generated content created by it would become a nightmare.
First. Please dont mass tag people x)
We dont like that
My apologies Kris
The issue of community management and collective intelligence operating content from & for many (introverts, public personalities and clients) is an issue that nobody talks about, we put in place an editorial process in place not an AI but we can consider it as a collectove general intelligence, hoever seems we have been considered as an AI which is not the case on hive.
Congratulations @jaxsonmurph! You have completed the following achievement on the Hive blockchain And have been rewarded with New badge(s)
Your next target is to reach 12000 upvotes.
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!
Very risky experiment indeed. Interesting however to see how quickly devs / gatekeepers / platforms pick up on the nuances. I noticed you literally yanked out a bunch of articles in a short period of time -- which is quite unusual (noticed it on Noise ;).
This being said, I prefer highly personalized and human-generated content. If we don't learn how to use these AI tools properly it's going to be a mess. BTW, I attended the Tech Show in London recently and listened to a few keynotes on AI, cloud storage and big data analysis. I saw an interesting tool that is being deployed in Vietnam currently to forecast dengue fever outbreaks in certain regions. Parties involved include but are not limited to medical doctors, meteorologists and government officials. That to me is a dandy use of AI.
Yep for sure, I'm the same I prefer the personalized engaging human content, but you are right if we don't come to grips with AI tech now given it's rather fast advances and the upsurge in investments it could turn into a disaster.
If used in the right way and by responsible people as you stated to track and assist with medical situations, that is a plus for humanity, but the risks of exploitation is troubling