Exploring conversational AI and poetry but not as we know it

2026-04-20 · Source: Artificial intelligence (AI) – The Conversation · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Social Sciences & Behavioral Studies, Content Creation & Production · Depth: Intermediate, short

Summary

Thierry Poibeau's new book, "Understanding Conversational AI" (Ubiquity Press), offers an interdisciplinary analysis of large language models (LLMs) and their impact on language, cognition, and society. The book explores how LLMs generate meaning, simulate reasoning, and perform complex tasks like translation, moral judgment, and literary creation, while also addressing their inherent limitations, biases, and roles in automation and misinformation. Poibeau delves into fundamental questions about understanding, creativity, agency, and trust in an era of synthetic language. Specifically, the book examines AI-generated poetry, noting its preference for quatrain structures, iambic meter, and end rhyme, often resulting in less nuanced emotional and metaphorical complexity compared to human-authored works. It also highlights how reader perception of AI poetry shifts once authorship is disclosed, despite initial comparable blind evaluations.

Key takeaway

For AI Ethicists and Research Scientists evaluating generative AI's societal impact, you should recognize that LLMs, while formally competent, often produce content that is stylistically homogeneous and less nuanced than human creations. Your assessment of AI-generated works, especially in creative fields, must account for evolving aesthetic norms and the significant influence of authorship disclosure on reader perception, rather than solely relying on initial blind evaluations. This necessitates a deeper consideration of authenticity and creative agency.

Key insights

LLMs challenge traditional notions of language, creativity, and authenticity, particularly in fields like poetry.

Principles

• Aesthetic judgment is historically contingent.
• AI poetry often favors accessible conventions.
• Disclosure impacts perception of AI-generated content.

In practice

• Evaluate AI-generated content with caution.
• Consider reader expertise in aesthetic judgments.

Topics

• Conversational AI
• Large Language Models
• AI-generated Poetry
• Aesthetic Judgment
• Natural Language Processing

Best for: AI Scientist, AI Ethicist, Research Scientist

Related on AIssential

• Beating the Style Detector: Three Hours of Agentic Research on the AI-Text Arms Race — Agentic research rapidly reproduces and extends NLP studies, demonstrating LLMs' ability to evade AI-text detection.
• LLMs Generate Kitsch — LLMs systematically generate kitsch due to their training, despite often outperforming human-generated works in ratings.
• The overusage of “It’s not A, it’s B” or “It’s not about A, but it’s about B” is driving me crazy. — AI-generated content's formulaic language and pervasive patterns are causing user fatigue and raising concerns about digital information integrity.
• I don't understand AI. How does it work? — LLMs predict text by mapping words to semantic vectors and learning statistical relationships from vast training data.
• How to Tell If Something Was Written by AI (And Why It’s Getting Harder) — Advanced LLMs now produce AI-generated content nearly indistinguishable from human writing.
• ‘Like a computer in my heart’: how AI poetry betrays our desire for human connection — Readers often prefer AI-generated poetry but feel betrayed upon discovering its artificial origin, highlighting the value of human connection.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Artificial intelligence (AI) – The Conversation.