AI’s Ode to an Ordinary Object

Is Chat GPT going to erode all of our trust in fellow poets?

As the fan in my room rattles and sqeals and I read through some of your competition entries, I wonder if I should write another. I then realise it’s 01:30 and I should be sleeping, but it’s a full moon tonight and I never sleep early on a a full moon.

Then I remember my trusty new friend ChatGPT and get curious about its poetry writing abilities.

“Write an Ode to a Fan” I instruct. Within seconds it is spitting out this poem onto my screen.

Ode to a Fan: A Whirring Whisper

Oh, humble fan, with blades that spin,

You bless us with a cooling din.

A whisper in the summer's heat,

Your gentle breeze, a sweet retreat.

.

From stagnant air, you set us free,

A loyal companion, it's you we see.

With each rotation, you unfurl,

A refreshing gust that makes us twirl.

.

In sweltering days, you work so hard,

A constant rhythm, like a bard.

Your whirring song, a soothing spell,

In your embrace, we find a swell.

.

You stir the air with grace and ease,

A guardian in the hottest breeze.

With gratitude, our hearts are full,

For your service, so dutiful.

.

You stand unnoticed, on the wall,

Yet, without you, we'd surely fall.

For in your presence, we're at ease,

A simple fan, a gift to seize.

.

Oh, blessed fan, we sing your praise,

In summertime's most sweltering days.

A treasure in the scorching weather,

You keep us cool, tied together.

.

So here's to you, dear fan of mine,

A constant presence, so divine.

In your embrace, we find our rest,

Ode to a fan, forever blessed.

If I saw this as an entry from a creator, I would believe it to be genuine. Before playing with this technology I would never have believed it was capable of doing this. It’s amazing.

The downside of course is that there are many pages of ‘creators’ now on Vocal who are using AI content to fill their pages. And they are not giving credit to it. Rather, they are plagiarising the content.

In the past 24 hours I’ve started copying poems into copyleaks.com (Thanks to Ashley Lima and her excellent article How To Spot AI) before wasting my time reading them. I don’t want to read plagiarised content. If I want to read AI content (and I do) I want to know I’m reading AI content.

I don’t want to be offering praise in the comments section thinking that I am helping to encourage a person and their creative abilities when actually I am only encouraging something wrong. In short, I don’t want my time and attention wasted on something that lacks purpose and meaning.

It is meaningful to support a human being into becoming positive, artistic and creative, for that is helping them into health. Health is the foundation of sustainable joy and true love. There is no other way.

Giving attention to people who are copying and pasting whilst being dishonest about it? That’s giving energy to something dark. The kind of world that we don’t want to grow.

I urge Vocal to take control of this. The problem is growing rapidly and many of the communities on here are now filled with AI generated spam and marketing. Real creators are being drowned in this sea of dishonesty.

Of course we are able to find real creators through the Top Stories feature, which I know we are all immensely grateful for. But what about the thousands who aren’t making it there? Whose writings are lost amongst an ocean of spam?

After reading Ashley’s article I decided to run the poetry I had recently liked and commented on through copy leaks. 6 out of 8 of them were 💯 AI generated. I felt cheated and disappointed, by both the ‘author’ and by Vocal not having checked this content for plagiarism. I scrolled down to my kind comments left to lift the heart of what I thought was a budding creator and promptly pressed the delete button.

This makes my heart sad.

I have created a second Vocal page to show us what the technology is capable of. I submitted to each Vocal Competition on behalf of ChatGPT so that you, as creators, are aware of what it can do. It’s intended to be conversational and fun. It makes it easier to spot the plagiarised content on here.

The problem is not the technology, or that it is being used to create articles. The problem is the dishonesty of people and the spamming of our communities on here. Perhaps there ought to be a community for AI generated content? Or strict rules that it must be labelled as such or accounts will be suspended?

Chat GPT, please offer advice to my favourite writing platform, Vocal, about how they can better manage the overwhelming amount of plagiarised content appearing on the site.

To help manage plagiarised content on Vocal, they can implement stricter content review processes using advanced plagiarism detection tools. Encouraging users to report suspicious content and establishing clear guidelines on originality will also be beneficial. Additionally, fostering a strong community that values creativity and unique voices can deter plagiarism and promote genuine contributions.

How can we detect AI content?

Detecting AI-generated content can be challenging, as AI models like GPT-3 can produce highly realistic and human-like text. However, here are some methods that can be used to identify AI-generated content:

1. Authorship Verification: AI-generated content may lack consistent writing styles or patterns compared to human authors. Analyzing the writing style, vocabulary, and syntax can sometimes help identify AI-generated text.

2. Sentiment and Emotional Analysis: AI-generated content might not exhibit genuine emotions or personal experiences. Analyzing the emotional tone and sentiment in the text can provide clues about its authenticity.

3. Pattern Recognition: AI models often generate content based on patterns found in their training data. Identifying specific patterns or phrases commonly used in AI-generated text could help in detection.

4. Response to Prompts: AI-generated content might overuse certain phrases when responding to specific prompts. Analyzing how the text responds to different input could reveal AI characteristics.

5. Turing Test: Evaluating the content through a Turing test or similar evaluation methods can help in distinguishing between AI-generated and human-written text.

6. Collaboration with AI Platforms: Partnering with AI platforms like OpenAI or GPT-3 to gain insights into potential AI-generated content and receive assistance in detection.

It's essential to combine multiple approaches and continuously improve detection methods, as AI models and their capabilities are constantly evolving.