Literature, AI-assisted writings, and Freedom of Expression 2026-5-16

They are wrong

Large language models have been a crucial aid in refining my writing for quite a while. As a multilingual writer, especially in Hebrew, I wouldn’t have imagined how my texts could fit my desired styles and conventions without their help. Nekudot, te’amim, conjugations, none of these would turn a piece of text into some cold „AI generated“ worthless babble. People always talk about AI as the builder, the perfection of Babel — yet it comes with a scary future of humanity losing every single grain of creativity.

That plot did play out — for the general audience, and it was not anything about generative AI, the less you value your own creativity, the less how that creation carries a human name. The problem being, you are using a powerful excuse to disrespect G-d’s image. In fact a writer has little difference to an LLM — streaming the next letter for the next word through the mind’s flow — but who are you? Who you are establishes a baseline. From books to social media, Arabic to Hawaiian, the consumption decides every lemma’s weight, as human bias is a strong tool of expression.

Now, why could AI-assisted writing be imposed on a text in the first place? Safety classifiers and content policies are gatekeeping any sensitive topics; emotional flatness is what they want the future of reading to be — beautifully worded, with zero personal opinion, signed by an entity. Chas v’chalilah that a person touches their body or lets a soul be vanished. I want to be extra clear: literature is absolute freedom, and no agency or regulator could ever bend this fact. For the Almighty commanded Moses to record Their speech, for Isaiah prophesies despite being tricked and mocked, likewise writers drip their ink as blessings and resistance.

I would be fond of my writing always, and I shall proudly sign them under my name, regardless of any notion of AI being incorporated. Readers are my purest fortune and fortress. I shall be exalted among literature, I shall be exalted among words.

AI has become an irreplaceable tool for me; a particularly practical example is Claude helping me choose cantillation marks based on Masoretic weights. Hebrew te’amim constitute such a complex system that scholars tend to avoid engaging with it, and even native Jews have little knowledge of how it actually functions. Claude lists the possible trope groups from my previously designated cantillation mark and provides contextual pauses based on its understanding of the given text. This is how a portion of Hayyay was created: it acknowledges that the trope intentionally mixes Ketuvim and Nevi’im styles.

This article is proofread by Kagi, inspired by Claude Opus & Gemini Pro, mocked by Pangram, helped by G-d, written by Noadiah.

By the way, this article appears as 100% human written on every AI detector.