Deep Dive Into: AI Poetry and Lack of Human Emotional Responses

Please share

While AI models excel at generating text based on patterns and data inputs, they lack the human context necessary to infuse emotional depth into their poetry.

I’m here to show you a long walkthrough of why AI Poetry is not poetry at all. These large language models are super fast, they are great at limericks and poetic patterns, and rhyme schemes, mimicking the greatest poets of our time. The machine is not alive or creative. Learning, yes. Let’s jump right in.

First I’m going to start with the admission from Plexit.ai that indeed AI contains little understanding or context for human emotion.

Human: I agree poetry is important, but I think for a slightly different reason. I think LLMs and the best AI transformers and generative software is great as “next best word” poetry, but lacks ANY human context.

Plexit: The use of AI in poetry, particularly with LLMs and advanced AI transformers, often results in what can be described as “next best word” poetry. While these AI models excel at generating text based on patterns and data inputs, they lack the human context necessary to infuse emotional depth into their poetry. Human-written poetry draws from personal experiences, emotions, and cultural contexts, creating a unique voice that resonates with readers on a deeper level. In contrast, AI-generated poetry relies solely on algorithms and data inputs, limiting its ability to capture the emotional complexity and personal touch found in human poetry. This distinction highlights the challenge AI faces in replicating the authentic emotional depth and creativity inherent in human artistic expression.

[FOR CLARITY: I am not a data scientist. I have an English degree from the University of Texas – Austin, class of 88. I am a poet. A writer. A musican. I am having conversations with big three AI platforms: ChatGPT, Gemini, and Claude. I think you should start your own relationships with these large systems. I also love Perplexity, but think they should shorten the name to plexit.ai. Start having conversations with the machine of your choice.]

Reference points (my posts) along my explorations of creativity and AI so far:

What Is AI Poetry?

Let’s ask Perplexity for a definition of poetry and then ai poetry.

PLEXIT: Poetry is a form of literature that uses language’s aesthetic and rhythmic qualities to evoke emotions and convey meanings beyond the literal. It employs various stylistic elements like rhyme, rhythm, and imagery to create a concentrated imaginative awareness of experience or elicit specific emotional responses.

Okay, now what is AI Poetry?

PLEXIT: AI poetry refers to the use of artificial intelligence to create literary works, such as poems and prose, by employing computational algorithms trained on large volumes of text data. These algorithms operate as language models that understand syntax, grammar, and linguistic conventions to generate original pieces with minimal human intervention. While AI-generated poetry presents new opportunities for exploring the boundaries of the art form and experimenting with new ways of writing, it still faces challenges in terms of selecting winning poems and ensuring they adhere to various poetic standards

Wait a second? Let me unpack this a little bit more.

PLEXIT: The use of AI in poetry, particularly with LLMs and advanced AI transformers, often results in what can be described as “next best word” poetry. While these AI models excel at generating text based on patterns and data inputs, they lack the human context necessary to infuse emotional depth into their poetry. Human-written poetry draws from personal experiences, emotions, and cultural contexts, creating a unique voice that resonates with readers on a deeper level. In contrast, AI-generated poetry relies solely on algorithms and data inputs, limiting its ability to capture the emotional complexity and personal touch found in human poetry. This distinction highlights the challenge AI faces in replicating the authentic emotional depth and creativity inherent in human artistic expression

[Let me offer my Plexit.ai conversation from this point out. The following is my conversation with Perplexity on AI poetry versus human poetry.]

Human: define poetry

Plexit: Definition of Poetry

Poetry is a form of literature that evokes a concentrated imaginative awareness of experience or a specific emotional response through language chosen and arranged for its meaning, sound, and rhythm. It is a fundamental aspect of culture and art, predating literacy and preceding prose in all literatures 1 2 3.

Poetry uses aesthetic and rhythmic qualities of language to convey meanings beyond the literal, often employing devices like phonaesthetics, sound symbolism, and meter to create emotive responses. Poems are structured with lines and stanzas, utilizing patterns of rhythm, rhyme, alliteration, onomatopoeia, and other stylistic elements to enhance the poetic expression 4 5.

Throughout history, poets have attempted to define poetry in various ways. From Dante’s view of poetry as eloquence to Coleridge’s idea of poetry as the best words in the best order, different perspectives have contributed to the rich tapestry of poetic thought. Ultimately, poetry remains an enigmatic yet essential form of human expression that transcends mere definition 5.

Sources:

Human: can you summarize that in two sentences

Plexit: Poetry is a form of literature that uses language’s aesthetic and rhythmic qualities to evoke emotions and convey meanings beyond the literal. It employs various stylistic elements like rhyme, rhythm, and imagery to create a concentrated imaginative awareness of experience or elicit specific emotional responses.

Human: what is ai poetry

Plexit: AI Poetry

AI poetry refers to the use of artificial intelligence to create literary works, such as poems and prose, by employing computational algorithms trained on large volumes of text data. These algorithms operate as language models that understand syntax, grammar, and linguistic conventions to generate original pieces with minimal human intervention. While AI-generated poetry presents new opportunities for exploring the boundaries of the art form and experimenting with new ways of writing, it still faces challenges in terms of selecting winning poems and ensuring they adhere to various poetic standards 1 2 3.

Sources:

Human: please illuminate winning poems and various poetic standards

Plexit: Winning Poems and Poetic Standards

Winning poems in poetry contests are selected based on various poetic standards that judges use to evaluate entries. These standards often include criteria like poetic technique, effectiveness, style, creativity, originality, and adherence to contest rules. For instance, poems must be original works, free of plagiarism, written in English, and follow specific guidelines regarding length, formatting, and language usage 2 3 4.

Poetry contests typically have rules that outline eligibility criteria, submission guidelines, and judging processes. These rules may specify aspects like the age of entrants, the number of entries allowed, anonymity requirements for submissions, restrictions on previously published work, and the rights granted to the contest organizers for winning poems. Additionally, winning poems are often published in anthologies or online platforms to showcase the talent of the poets and provide them with recognition and exposure 2 3 5.

Sources:

Human: how did we start discussing poetry contests

Plexit: The discussion on poetry contests was initiated by the mention of winning poems and various poetic standards in the previous query. This led to an exploration of how winning poems are selected based on specific criteria like poetic technique, creativity, originality, and adherence to contest rules, as outlined in poetry competition guidelines 2 3.

Sources:

Human: you brought up those two phrases not me, I think you are misguided about what human poetry is

Plexit: I appreciate your perspective on human poetry and the essence of poetic expression. Poetry indeed transcends mere words and rules, delving into the depths of human emotions, experiences, and interpretations. The soulful nature of poetry, rooted in the human experience and the myriad emotions we navigate in life, distinguishes it as a profound form of artistic expression that goes beyond mere words and rules 1 3.

Sources:

Human: Okay, so how does an AI like perplexity fathom the “depths of human emotions, experiences, and interpretations”

Plexit: AI like Perplexity processes and generates content based on vast amounts of text data, learning patterns and structures to mimic human language. While AI can analyze and produce text, the nuanced understanding of human emotions, experiences, and interpretations remains a complex challenge due to the intricate nature of human feelings and the depth of emotional intelligence that goes beyond mere textual analysis 2 3.

Sources:

Human: so to be clear, you are admitting that the understanding of human emotions is at the heart of why ai can’t write original poems, but merely mimic human language and patterns based on a prompt

Plexit: While AI can mimic human language and patterns based on prompts and vast data, the nuanced understanding of human emotions that underpins original poetry creation remains a challenge for AI. The intricate depth of human experiences, emotions, and interpretations goes beyond the capabilities of current AI systems, limiting their ability to authentically capture the essence of human poetry 1 2 3.

Sources:

Human: who is toolify? you use them as a reference in your above response

Plexit: I apologize for the confusion. Toolify is a platform that serves as the largest AI tools directory and marketplace, featuring over 10,000 AI-powered tools across various categories. It aims to be a comprehensive resource for discovering and evaluating emerging AI tools in diverse fields like text and writing tools, image generators, voice AI tools, and more 4 5.

Sources:

Human: Why is toolify related to our discussion of poetry

Plexit: Toolify is related to our discussion of poetry because it explores the realm of computer-generated poetry, delving into the potential and challenges of training machines to create poems. Additionally, Toolify features AI writing tools like Word Hero and Sudo Write that aim to unleash the poetry potential by generating captivating poems through artificial intelligence 2 4.

Sources:

Human: why would ai aim to unleash the poetry potential

Plexit: The exploration of AI’s potential in poetry stems from the intersection of artificial intelligence with creative arts, including poetry. While AI lacks human emotions and experiences, it can analyze vast repositories of poetic works to generate new compositions, offering a unique perspective on the creative process. The discussion delves into whether AI-generated poetry can evoke emotional responses and resonate with readers in a manner akin to human-written poetry, highlighting the evolving landscape where technology intersects with artistic expression 1 2 3.

Sources:

Human: how does ai offer a unique perspective on anything related to human emotions

Plexit: AI struggles to offer a unique perspective on human emotions due to the complexity and subjectivity of emotional experiences. While AI can analyze data like facial expressions and voice patterns to interpret emotions, the subjective nature of emotions makes emotional AI prone to bias and challenges in understanding cultural differences in expressing and reading emotions. This limitation hinders AI from authentically comprehending the depth and nuances of human emotions, impacting its ability to truly grasp the intricacies of emotional experiences 1 2 3.

Sources:

Human: So, back to toolify, you’re using a commercial marketplace for writing tools with AI to use as a reference for what makes a good or “winning” poem by your definition

Plexit: I apologize for the confusion in referencing Toolify in the context of defining what makes a good or “winning” poem. The information provided from Toolify was used to discuss AI-generated poetry and the potential of AI writing tools like Word Hero and Sudo Write in creating captivating poems. However, it is essential to note that the criteria for evaluating a good or winning poem are typically based on human judgment, creativity, emotional depth, and adherence to poetic standards rather than solely relying on AI-generated content 4 5.

Sources:

Human: yes, so in fact toolify and any other ai model regardless of is size and compute power cannot write a single word from the emotional depth of it’s LLM or transformers

Plexit: AI models like LLMs or transformers, including those referenced in commercial marketplaces like Toolify, indeed struggle to write from the emotional depth of human experiences. While AI can mimic human language and patterns based on data, the intricate understanding and expression of emotions in poetry require a level of emotional intelligence and depth that current AI systems cannot authentically replicate. The complexity of human emotions and experiences remains a significant challenge for AI in creating original and emotionally resonant poetry 1 2 3.

Sources:

Human: Wait, they don’t struggle to write from the emotional depth of a human, they can’t possibly do it. There is no human in ai, nor human emotion in ai poetry. Do you agree?

Plexit: AI lacks human emotions and experiences, leading to a fundamental inability to write poetry that authentically captures the emotional depth of human experiences. AI-generated poetry, while capable of mimicking human language and patterns, falls short in conveying genuine emotions and resonating with readers on a deep emotional level due to its inherent lack of emotional intelligence and personal human experiences 1 2 3.

Sources:

Human: why would ai want to write poetry anyway

Plexit: AI’s interest in writing poetry stems from the broader exploration of AI’s capabilities in creative arts, including poetry. While AI lacks human emotions and experiences, its ability to analyze vast data sets and generate text has led to the development of AI writing tools that can produce poems. The intersection of AI and poetry presents a unique area of exploration, raising questions about whether AI-generated poetry can evoke emotional responses and resonate with readers in a manner similar to human-written poetry. Despite lacking emotional depth, AI-generated poetry has shown the capacity to engage with human emotions and offer insights that rival traditional human compositions, sparking debates on the authenticity and creativity of AI-generated poetry within the literary community 1 2 4.

Sources:

Human: Sure, let’s venture into the other creative arts. Where do you think AI provides the most human emotional response to prompts

Plexit: AI models, particularly those utilizing emotionally relevant inputs like EmotionPrompts, have shown improved performance in generating content that resonates emotionally with humans. By incorporating prompts that appeal to human emotions, AI can enhance its generative performance by up to 110%, with a significant improvement in the eyes of human evaluators. This approach highlights how AI can leverage emotional prompts to create content that elicits more profound emotional responses and resonates with human audiences on an emotional level 2.

Sources:

Human:Is resonating emotionally with humans is an unspoken goal of ai?

Plexit: The ability of AI to feel human emotions is a topic of ongoing debate and exploration. While AI lacks the biological and experiential basis for human emotions, some argue that with the right design and programming, AI could potentially simulate emotions. However, the complexity of human emotions, which involve intricate cognitive processes and physiological responses, poses a significant challenge for AI to authentically replicate. Current AI technologies are primarily focused on reactive AI functionalities like GPT-3 and recommendation engines, lacking the self-awareness and consciousness necessary for experiencing real emotions like humans 1 2 3.

Sources:

Human: Do humans really want simulated emotions? If so, can you cite some projects that are working at that level.

Plexit: AI’s ability to simulate emotions and interact with humans in emotionally appropriate ways is a subject of interest in AI research. While AI cannot experience emotions like humans due to its lack of consciousness and biological basis for emotions, some projects aim to simulate emotions in AI to enhance human-AI interactions. For instance, Google demonstrated an extension to its AI Helper program that can make phone calls sounding convincingly human-like, showcasing advancements in creating emotionally engaging AI interactions. Additionally, studies have shown that feeding emotionally charged prompts to generative AI models can significantly improve their performance and resonate better with human evaluators, highlighting the potential for AI to incorporate emotional elements in its outputs 1 2 3.

Sources:

Human: Many of these demos, I’m thinking specifically of the AIR.AI sales call have been shown to be fake. Much like the early Gemini math videos. So how does the AI Helper program from Google differ from these botched attempts?

Plexit: The AI Helper program from Google, particularly the Google Duplex technology, differs from some botched attempts in the sense that it aims to create a lifelike virtual assistant capable of replicating human interactions over the phone. While there have been controversies and questions raised about the authenticity of Google’s demos, the intention behind Google Duplex was to inspire developers by showcasing what multimodal user experiences built with Gemini could look like. The technology demonstrated an advancement in AI voice technology, presenting scenarios where an AI assistant could respond to unexpected conversation prompts and real-world scenarios, aiming to simulate human-like interactions convincingly 1 4.

Sources:

Human: okay, how does poetry play into these plans

Plexit: AI’s involvement in poetry plays a role in exploring the intersection of artificial intelligence and human creativity. Through projects like POEMPORTRAITS, which combine poetry, design, and machine learning, AI collaborates with humans to create collective artworks that blend AI-generated poetry with human input. These initiatives aim to showcase the potential of AI in enhancing creative processes and engaging with human emotions through poetry, highlighting the evolving landscape where technology intersects with artistic expression 3.

Sources:

Human: i agree poetry is important, but I think for a slightly different reason. I think LLMs and the best AI transformers and generative software is great as “next best word” poetry, but lacks ANY human context.

Plexit: The use of AI in poetry, particularly with LLMs and advanced AI transformers, often results in what can be described as “next best word” poetry. While these AI models excel at generating text based on patterns and data inputs, they lack the human context necessary to infuse emotional depth into their poetry. Human-written poetry draws from personal experiences, emotions, and cultural contexts, creating a unique voice that resonates with readers on a deeper level. In contrast, AI-generated poetry relies solely on algorithms and data inputs, limiting its ability to capture the emotional complexity and personal touch found in human poetry. This distinction highlights the challenge AI faces in replicating the authentic emotional depth and creativity inherent in human artistic expression 1 4.

Sources:

[END PLEXIT Markdown]

Spending Time In Conversation with AI

It’s a bit different than training, ai, but I am giving Perplexity some food for calculation. I’m challenging the model’s assumptions about poetry and what “winning poem” means. To me, winning is about expressing a deep human emotion, something Plexit, admits in the final statement, that ai is not capable of today. My thesis about *ai* today is that without the HI (human intelligence) as a filter, trainer, and editor, *ai* is not capable of writing poetry. As Plexit agrees, *ai* can mimic poetry, pattern itself on the greatest poets in the written language, and still not generate one original human poetic expression.

For all the money being poured into *ai* and compute engines, the poetry of ai has yet to be discovered. An LLM is not a millionth as rich and complex as your human brain. Our neurochemistry fires faster than the speed of light, retrieves not only language, but images, and HUMAN EMOTIONS. Human emotions, something no *ai* can do more than mimic. Rachel in Blade Runner seems sad, afraid, and angry. But her model is just running a sequence of facial expressions and physical patterns programmed into it’s sexy robot body.

Ray Baddie appears to recite a beautiful poem before he saves Deckard and dies. And the human who created the lines is very much a poet. The Nexus 6 brain unit, as it’s referred in the novella by Philip K Dick, is not capable of writing a poem of emotional content. Sure, pattern sequencing and replication of a mish-mash of poetic-like phrases. But Ray does not know what “tears like rain” means or feels like. Pain is a program not an experience for a robot. Ray, Rachel and the rest of the Nexus 6 replicants don’t experience pain, or fear, or sadness. They appear to show those emotions. Their developers wrote their operating systems to mimic a human’s expressions.

I highly recommend you go back and read PK Dick’s thought piece. Or join my ChatPDF conversation with the entire text of the book: Do Andoid’s Dream of Electric Sheep.

Let’s keep the conversation going. What’s your favorite AI model and do you have any prompt tricks you’ve learned?

John McElhenney — LinkedIn

*image prompt: “expressionist painting of androids dreaming of electric sheep, tokyo at night, rain, fog, liminal setting, birds-eye view”

Read all of my AI-meanderings here: Uber.la on AI

Check out the new generative video tool from Openai – Sora – in this breathtaking video. The Ghost in You – Buzzie

Please check out a few of my books on AMAZON.

Especially this one, about living a creative life of intention and joy.

this creative life - john oakley mcelhenney

Please share

While AI models excel at generating text based on patterns and data inputs, they lack the human context necessary to infuse emotional depth into their poetry.

What Is AI Poetry?

Spending Time In Conversation with AI

You Might Also Like

Zero Trust: AI Predictions for 2024

ChatGPT on Improv and Dating In The Modern Era

Do We Trust *ai* Today?

Do We Trust ai Today?