AI
Authors Turn to 'Process Tracking' Tools to Prove Their Work Is Human—But Experts Question If It Works
As AI detection accusations roil the publishing world, new software platforms promise to document writers' creative workflows—yet copyright experts warn the technology is far from foolproof.
WIRED — AI
Publishing faces a novel crisis: authors facing public allegations that their work was generated by artificial intelligence, with few reliable ways to defend themselves. In a landscape of growing AI detection tools and reader skepticism, writers are turning to a new class of software that aims to create verifiable proof of human authorship.
The Scandal Cycle Emerges#
The publishing industry confronted this dilemma head-on last March when Shy Girl, a horror novel, became the subject of widespread social media debate over its AI origins. After a New York Times article examined those claims, Hachette Book Group pulled the book from its upcoming US release and discontinued the UK edition. According to the AI detection platform Pangram, roughly 78 percent of the text showed signs of artificial generation.
Author Mia Ballard has contested the allegations, asserting that an editor—not she—introduced AI-assisted revisions to a previously self-published manuscript. Since then, additional fiction writers have faced similar accusations, including one who won a short story prize and another whose crime novel commanded millions in publisher interest. All have denied the charges.
The Trust Problem#
The Shy Girl episode has created what novelist Andrea Bartz—lead plaintiff in a landmark copyright lawsuit against Anthropic—describes as a potential "witch hunt." Authors now fear that unsubstantiated claims of AI usage could spread rapidly among readers, regardless of evidence. Bartz noted that she has already witnessed review-bombing campaigns on Goodreads and readers pledging to avoid books published after 2022 to ensure authenticity.
The core issue: there is no standardized method to establish human authorship. Once accusations surface, the reputational damage is often irreversible, leaving writers exposed to public judgment based on subjective impressions and imperfect detection algorithms.
New Tools Promise Documentation#
A constellation of platforms—including Workings, Blindproof, Wintertext, Human Ink, and ObraVera—now offer to track and timestamp an author's writing process. Some encrypt and fingerprint typing patterns or document progression through multiple drafts, creating what proponents call a "paper trail" of authorship. The goal extends beyond defense against accusation: these systems could also establish provenance if text is later scraped into AI training sets without permission or proper attribution.

