The Scarlet Letter for the Honest: A Short Pamphlet on the New Witch Hunt

The pamphlet argues that AI labels can punish ordinary, transparent use of writing tools while failing to identify industrially generated content. It distinguishes assistance, where a person makes the substantive decisions and uses software for tasks such as editing or translation; generation, where a human supplies an idea but a model produces the wording; and mass production, where texts are published with little or no human review. A single “AI-processed” badge erases these differences. It can attach to a writer who carefully checked a machine-assisted translation, yet disappear from generated material after simple paraphrasing. The label therefore measures contact with a tool, not the author’s role or responsibility.

The pamphlet also challenges AI detectors as unreliable, arguing that they identify polished, neutral, well-structured prose rather than machine origin. It cites a pre-LLM text that received radically different scores at different times and notes that canonical writing can be falsely flagged, while distinctive personal style may escape detection. This turns competent writing into suspicious evidence and shifts the burden onto authors to prove they are human. The same double standard appears when readers reject visibly labeled text but accept unmarked AI search summaries. The proposed remedy is a presumption of human origin unless reliable, verifiable evidence shows otherwise; probabilistic detectors may prompt investigation but should not justify penalties. The decisive question, it concludes, is who made and checked the intellectual decisions behind a text, not what its wording looks like.