← Encyclopedia Skill

AI in Poetry: Disclosure, Originality, and Verification

Introduction Does a poem fully generated by artificial intelligence become "your poetry" simply because you publish it under your name with no meaningful edit at all? And does merely using a tool that suggests a rhyme or checks a meter require the exact same disclosure that full generation requires? The answer is neither an absolute no nor an absolute yes for every undifferentiated use — it depends precisely on the actual scale of your contribution to the text and the context in which the final work is published. Definition and distinction There is a fundamental difference that should not be ignored between using an AI tool as an assistant that suggests alternative images, reviews prosodic meter, or proposes possible rhymes — which you then select, revise by hand, and weave into a text you essentially wrote yourself — and asking the tool to generate a complete poem and publishing it exactly as returned, with no substantive edit, under a human name attributed to you. The first is using a tool much like a dictionary or an advanced spell-checker that assists rather than replaces; the second is a full, effective replacement of the creative writing process itself. Two genuinely different contexts also call for very different levels of disclosure: personal practice and training, where you experiment with the tool to learn or explore new ideas without the output being published or evaluated by anyone else, needing no disclosure whatsoever; and work that is formally published, evaluated, or entered into a literary competition, where clarity about the tool's role becomes a genuine part of the process's own integrity, because the judgment here falls on the person's real skill, not on the quality of the final output alone. A realistic scenario Picture a poet who uses an AI tool to suggest five alternative images for the phrase "a sad sunset," chooses one of them, revises it substantially, and weaves it into a poem she wrote herself at well over ninety percent of the total text. Compare this to another poet who directly asks the tool to "write a complete poem about longing in the basit meter" and publishes it exactly as returned, with no edit at all, under his own personal name, in a formal literary competition that judges individual skill. The first used the tool as a limited, partial source of inspiration within work that was fundamentally her own and produced through her real effort; the second published the tool's complete, unedited output, fully attributed to his personal skill, with no acknowledgment of any kind of the tool's actual role in producing it. Practical application Step 1: Honestly assess the actual ratio of your contribution — the original idea, the final wording, the substantive revision — against what the tool generated with no meaningful edit at all. This honest internal assessment is the basis on which you build your subsequent decision about disclosure. Step 2: If you publish the work in any evaluative, competitive, or commercial context, disclose the tool's role in a short, honest sentence — noting, for instance, that some images were suggested with a tool's help and then fully reworded by hand — rather than staying completely silent about it or denying it outright when asked directly. Step 3: Manually and carefully verify any "fact," citation, or line attribution the tool suggests before using it in your work, because AI tools can fabricate lines that do not exist at all or attribute them to entirely the wrong poet with a confident tone that sounds convincing to the reader despite being fundamentally wrong. Step 4: Do not use the tool to imitate a specific living contemporary poet's style to a degree that creates genuine confusion about the text's identity for the reader, without an explicit, clear note that the work is a deliberate, intentional pastiche rather than an independent original creation in its own right. Context that changes the meaning The Zayenha Arts platform itself uses artificial intelligence to evaluate certain artistic skills, and this absolutely does not mean every undifferentiated use of the tool in actually producing poetry itself is acceptable regardless of the case; there is a fundamental difference between an evaluation tool that reveals a skill level the user genuinely already possesses, and a generation tool that can produce text falsely attributed to a skill that has not actually been developed in reality yet. Honest disclosure is precisely what preserves the credibility of both systems at once, with no real contradiction between them. Educational contexts, such as training students learning the basics of meter and rhyme for the first time, generally allow broader experimentation with the tool as long as the primary goal is personal learning and exploration rather than final publication fully credited to the individual student's own skill. Competitive and evaluative contexts, such as formal literary competitions or accredited skill-assessment tests, require far clearer disclosure than that, because judging a person's own "skill" is the very point of the entire exercise, and a dishonest result in this context directly wrongs every other participant who relied solely on their own personal effort without outside assistance. Limits of the term and common mistakes A common mistake is assuming that any technical assistance at all, even simple spell-checking or a single suggested synonym for a word, is exactly equivalent to full generation requiring the very same disclosure — an imprecise generalization that conflates entirely different levels of actual reliance on the tool. A far more serious mistake is denying use of the tool when asked directly and explicitly, which turns a purely technical question into a matter of fundamental personal integrity. Another common mistake is placing absolute trust in lines or citations the tool suggests without any real verification of a genuinely trustworthy source, on the mistaken assumption that a tool's confident phrasing necessarily means the suggested content is accurate. Finally, it is a mistake to believe that using the tool only at the earliest brainstorming stage, with no trace of it left in the final text, still requires disclosure; the real standard is the final effect on the text itself, not every single step the writing process happened to pass through. Conclusion The practical standard: if the tool's contribution were entirely removed from the poem and a coherent text remained that genuinely and honestly expressed you, your use is generally legitimate and needs no lengthy, detailed disclosure. If the text collapses or loses its coherence entirely once that contribution is removed, disclose its real role clearly before anyone else has to ask you about it themselves.