Tuesday, 17 December 2024
10 ways to survive Invisible Poets
Friday, 19 July 2024
Why AI Can't be Creative
Geometric city. Anonymous streetsLabelled with letters and numbers????? the shops into blocks.
Wednesday, 10 April 2024
AI: Threat or opportunity?
AI is nothing new. I first started experimenting with machine learning or artificial intelligence (AI) way back in the early 1980s. The phrase 'Artificial Intelligence' had been coined way back in 1955 by John McCarthy, but many years before this, others had been speculating on, and experimenting with the idea that machines could 'think' or at least mimic human patterns of thought. The entire history of AI can be read at this link.
In 1966 computer scientist Joseph Weizenbaum developed a program that mimicked a psychotherapist. He called it Eliza, and it was probably the first artificial dialogue program (or chatbot) ever created.
I sit at the nexus between psychology and computer science. As a former associate professor in education I have a professional and personal interest. In 1982 I developed a chatbot based on Eliza's source code which I named Dr Fraud. Essentially, the program analysed whatever you inputted and then spat back an insult or some form of abuse at you. This rapidly became a very popular program with my student nurses (especially the psychiatric nurses), and they queued up to use it. Often you would hear gales of laughter as the student nurses read the rude, but inventive lines the 'psychiatric insultant' chatbot known as Dr Fraud would hurl at them. It was all good fun, and it introduced my students to the idea that not only could they learn from using computers, but that it could be fun, because of the dialogic and interactive elements.
Today of course, AI has developed rapidly from the early forays to the point that it is now a constant news and media item, and everyone seems to be talking about it. Generative AI (G-AI) is built into so many applications we use, from banking and shopping to home use of devices such as Siri, Alexa and smartphones. The popularity of free AI tools such as ChatGPT, DALL-E and Synthesia is transforming the way we create, analyse and disseminate knowledge and content.
Where poetry is concerned, we can perceive AI as either a threat or an opportunity. Most online poetry groups have banned the sharing of AI generated poetry because in essence it is plagiarism. The AI tool will spider the web for existing poetry texts, steal from them and construct a 'poem' for the user. It takes away from the creativity of poetry and makes a mockery of the process of writing a poem. Why claim to be a poet when all you are doing is pressing a computer key? There are similar arguments from the art, entertainment, literary and music industries who sense the same threat.
What is the alternative perspective? It is this - that in all of the above fields, AI can be used as a starting point, an inspiration or a series of steps toward creating something new and unique. The view is that creativity can be supported and even enhanced through the use of G-AI tools, if used sensibly, ethically and appropriately.
So for example, could AI be used to prompt an idea for a poetry topic, or to offer a template for a villanelle or a pantoum? Could it be used to suggest s good end rhyme, or as a check for grammatical or syntactic accuracy. How about using it to translate your poem into another language... or as a co-author of your work. Which of these is ok and which is going to far?
The jury is still out. What are your views on the use of G-AI in poetry?
Steve Wheeler
Image from Wikimedia Commons
Friday, 26 January 2024
How to write mindless poetry
Do you want to write mindless poetry? Easy. Get a machine to do it. You won't need to think about it at all. You simply press the key and the computer will write your poem for you. Sadly, this is becoming a popular trend on Facebook poetry groups.
Several people have asked me why we ban the use of Artificial Intelligence (AI) generated poetry on Invisible Poets group. It's quite simple.
Firstly, it's very lazy. Anyone can click on ChatGPT for example, and ask it to generate a poem on any subject (many other free AI tools are available). This requires absolutely no talent or creativity. Let's say you want a five stanza poem on love. In seconds you have something like this:
In the dance of hearts, a melody sweet,
Love's symphony, in whispers discreet.
Through moonlit nights and sunlit days,
A timeless bond that forever sways.
Eyes that speak in a silent prose,
A language only true lovers know.
In every touch, a tender embrace,
Love's magic transcends time and space.
Through stormy seas or skies serene,
Love's fortress strong, an eternal sheen.
A tapestry woven with threads so fine,
Two souls entwined, an intertwining line.
In laughter shared and tears that fall,
Love weathers storms, standing tall.
A journey marked by moments divine,
A love story, forever to shine.
So let the echoes of love resound,
In every heartbeat, in love, be found.
A poem written in the language of the heart,
A masterpiece, where love is art.
Well, at first glance, it seems like a reasonable poem. But read it more thoughtfully and you'll see it's not very good poetry. It is literally mindless. It uses cliches in a predictable 8 syllable, 4 step AABB rhyme scheme. It's bland, soulless, and takes no creativity or effort on your part. In ChatGPT-3, the same words seem to appear repeatedly, such as serene, tapestry, entwined, embrace, eternal, etc. These are all tell-tale signs the poem was not written by a human.
Now why would you want to put your name to such drivel?
Furthermore, you wouldn't do it if you knew how AI actually works. It spiders (searches) all the poetry on the web (other poets' work), gathers expressions, lines and rhymes and stitches it all together so you can pretend you wrote it. It's plagiarism. It's dishonest. It's theft.
Now why would you want to associate yourself with that? And yet... and yet... a small minority in the group still persist, thinking they can fool everyone. They end up being banned from the group, and their name forwarded to the Admin of many other groups. Game over.
If we suspect a poem has been AI generated we remove it from the group. If the member persists and posts more, we eventually remove them from the group. We throw people out of the group every week for trying to pass of AI poetry as their own. We don't lose sleep over it. We want to protect the integrity of the group and the interests of our members. There is software that can detect whether AI has been used, and we also have computer scientists on our Admin/Moderator team who are experienced in the use of AI.
So please don't even think about generating poems using AI. Write your own poetry. Poetry you can be proud of. Express yourself in your own words. Steer away from the kind of bland, soulless AI lines you see in the verses above. We don't want mindless poetry. We want soulful, authentic, from the heart poetry!
Come back at me if you disagree with any of this. I like a good fight.
Steve Wheeler
Image from Pxhere
Pushing the Boundaries
Yesterday I was in the studio recording a series of short radio shows in my Poets Corner slot for CrossRhythms Radio . The show is divided i...
-
Do you want to be a part of something really great? Something that is far more than just poetry? Something that will leave a lasting legacy ...
-
Do you want to be a part of something really great? Something that is far more than just poetry? Something that will leave a lasting legacy ...
-
With two new anthologies about to be published by Wheelsong Books , I'm sure there will be many in the groups who would like to know how...



