Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Creating a digital poet

Vered Tohar, Tsahi Hayat, Amir Leshem

TL;DR

The paper investigates whether a machine can write high-quality poetry and shows how a large language model can be guided to develop a coherent, distinctive poetic voice through extended, workshop-style prompting with expert feedback, without any parameter updates. Using a seven-month protocol of principle-led drafting, critique, revision, and rule distillation, the authors demonstrate long-horizon in-context learning that yields a stable persona and a coherent, expandable poetry corpus. In a blinded authorship test with humanities students and graduates, readers judged AI and human poems at chance, suggesting the model's outputs can closely resemble human poetry when evaluated in isolation. Separately, the model generated a publishable poetry collection, illustrating a path from workshop process to professional publication and prompting discussions about authorship, originality, and creativity in AI-assisted art, while noting limitations in rhyme and meter control and the influence of trainer priors.

Abstract

Can a machine write good poetry? Any positive answer raises fundamental questions about the nature and value of art. We report a seven-month poetry workshop in which a large language model was shaped into a digital poet through iterative in-context expert feedback, without retraining. Across sessions, the model developed a distinctive style and a coherent corpus, supported by quantitative and qualitative analyses, and it produced a pen name and author image. In a blinded authorship test with 50 humanities students and graduates (three AI poems and three poems by well-known poets each), judgments were at chance: human poems were labeled human 54% of the time and AI poems 52%, with 95% confidence intervals including 50%. After the workshop, a commercial publisher released a poetry collection authored by the model. These results show that workshop-style prompting can support long-horizon creative shaping and renew debates on creativity and authorship.

Creating a digital poet

TL;DR

The paper investigates whether a machine can write high-quality poetry and shows how a large language model can be guided to develop a coherent, distinctive poetic voice through extended, workshop-style prompting with expert feedback, without any parameter updates. Using a seven-month protocol of principle-led drafting, critique, revision, and rule distillation, the authors demonstrate long-horizon in-context learning that yields a stable persona and a coherent, expandable poetry corpus. In a blinded authorship test with humanities students and graduates, readers judged AI and human poems at chance, suggesting the model's outputs can closely resemble human poetry when evaluated in isolation. Separately, the model generated a publishable poetry collection, illustrating a path from workshop process to professional publication and prompting discussions about authorship, originality, and creativity in AI-assisted art, while noting limitations in rhyme and meter control and the influence of trainer priors.

Abstract

Can a machine write good poetry? Any positive answer raises fundamental questions about the nature and value of art. We report a seven-month poetry workshop in which a large language model was shaped into a digital poet through iterative in-context expert feedback, without retraining. Across sessions, the model developed a distinctive style and a coherent corpus, supported by quantitative and qualitative analyses, and it produced a pen name and author image. In a blinded authorship test with 50 humanities students and graduates (three AI poems and three poems by well-known poets each), judgments were at chance: human poems were labeled human 54% of the time and AI poems 52%, with 95% confidence intervals including 50%. After the workshop, a commercial publisher released a poetry collection authored by the model. These results show that workshop-style prompting can support long-horizon creative shaping and renew debates on creativity and authorship.
Paper Structure (2 sections, 8 equations, 4 figures, 2 tables)

This paper contains 2 sections, 8 equations, 4 figures, 2 tables.

Table of Contents

  1. Workshop protocol and curriculum.
  2. The “Naomi Efron” persona.

Figures (4)

  • Figure 1: Teaching with expert feedback. Prior session knowledge initializes the context for new poem generation. Outputs are evaluated and iteratively revised until validated, then consolidated into persistent memory. Session summaries update contextual knowledge, enabling cumulative improvement across interactions.
  • Figure 2: Naomi Efron. The model was asked to generate an image of itself. The presented image is the response to this prompt.
  • Figure 3: Experiment design and main result. Subfigure (A) presents the experimental process. (B) presents the proportion of identification of human writen songs and model-generated songs as human. 95% confidence intervals are also presented.
  • Figure 4: Distribution of subject-level accuracies across the 50 valid subjects. The dashed vertical line marks chance-level accuracy (0.5).