Certified into the record
The exact text we saved
This is the version exactly as it was saved, character for character. The SHA-256 hash below is a 64-character code worked out from that text. Run the same text through SHA-256 and you get the same code; change a single letter and the code comes out completely different. Each version also stores the previous version’s hash, which chains them together so no version can be swapped out quietly.
TRUVACE RECORD VERSION record: TRV-2026-2916 version: 1 kind: certified reason: Certified into the record timestamp: 2026-11-14T18:01:43.056126Z status: published lens: p_space sector: entertainment headline: ‘It’s like the wild west’: how AI has rocked the literary world dek: The romance novelist Coral Hart got her break writing books for Mills & Boon, the publisher of bodice-ripper paperbacks. In 2012, she decided to dip her toe in self-publishing. She felt drawn to dark romance, a subgenre featuring twisted, toxic love stories that was taking off online but that publishers didn’t want to touch. Three years later, things were going well enough that she decided to self-publish full-time. Kindle Unlimited, where Hart sells her books, rewards prolific writers: the algorithm favours… gain_title: (none) problem_title: Since March 2026, Hachette pulled the novel Shy Girl, agents dropped a $2m deal and a novel was struck off the Goncourt longlist over AI-use accusations that authors denied or disputed, with no definitive proof either way. trace_subject: (none) gain_reading: (none) gain_evidence: (none) problem_reading: Since March 2026, Hachette pulled the novel Shy Girl, agents dropped a $2m deal and a novel was struck off the Goncourt longlist over AI-use accusations that authors denied or disputed, with no definitive proof either way. problem_evidence: Hachette withdrew the romance novel Shy Girl from publication over evidence the author Mia Ballard used AI | There is no definitive proof either way. quick_read: Publishing has seen a run of cancelled book deals and accusations over AI use in 2026. In March Hachette withdrew the romance novel Shy Girl, in July agents dropped debut novelist Jerry Falade's $2m deal, and last month Thélyson Orélien's novel was struck off the Goncourt Academy longlist after social media users ran it through AI detection software. The accused authors have mostly denied using AI to write, and several declined to speak on record. The article notes that every writer at the centre of these scandals has been an author of colour, and that non-native and neurodivergent writers are more likely to draw false suspicion. Meanwhile AI is generating real income and anger. Romance author Coral Hart says she has written 370 novels under 21 pen names since 2025, earning 'high six figures', while Anthropic paid a $1.5bn settlement in a class action covering around 500,000 titles. An industry body committee says agents disagree on whether to use detectors routinely or rely on trust. limitation: Most accused authors deny using AI and the article says there is no definitive proof either way, so the accusations remain unproven in several cases. tag: Evidence-backed problem key_points: Hachette withdrew the romance novel Shy Girl in March over evidence that author Mia Ballard used AI; she said an editor she hired may have inserted AI text. | Agents withdrew from a $2m book deal with debut novelist Jerry Falade in July; he denied using AI beyond research. | Orélien's novel was struck off the Goncourt longlist and his publisher Boréal dropped him in late October; he denied using AI but admitted plagiarism in other writing. | The 2026 Commonwealth short story prize winner Jamir Nazir was accused after readers ran his story through Pangram; he denies it, and Pangram's checks flagged four regional winners since 2025 as likely AI users. | Academic research estimates Pangram is at least 95% accurate with a near-0% false positive rate, yet human judgment is flawed and non-native and neurodivergent writers are more likely to be wrongly suspected. | A preprint led by Tuhin Chakrabarty of Stony Brook estimates about one fifth of recently self-published books contain at least 25% AI-generated text. | Anthropic won a $1.5bn settlement in a class action involving around 500,000 titles; 70 authors asked the big five publishers to promise never to release machine-made books, and the publishers did not. rundown: Coral Hart, who started at Mills & Boon and moved to self-publishing, says AI let her write 370 novels under 21 pen names since 2025. Because Kindle Unlimited pays per page read, she argues readers abandon bad books, so craft still matters. She now teaches AI book-production courses and advises authors not to disclose AI use unless required, saying disclosure brings death threats; she was doxed after speaking to the New York Times. The cases of Falade, Nazir and Orélien began largely when readers or journalists ran work through detectors. Falade wrote that non-native or culturally unfamiliar English is more often treated with suspicion and suggested racism played a part. Author Steven Rosenbaum, who is white, disclosed AI use in his book The Future of Truth, which contained fabricated quotes; his book was not withdrawn. Hannah Bowman, chair of the Association of American Literary Agents' AI special committee, says agents disagree: some want detection software used routinely, others prefer trust. She wants a fair process in which authors can respond to concerns. Book scout Liz Gately says agents now advise writers to avoid sending PDFs and to save multiple drafts in case they face accusations. sources: - journalism | The Guardian | https://www.theguardian.com/books/ng-interactive/2026/nov/14/its-like-the-wild-west-how-ai-has-rocked-the-literary-world | 2026-11-14 prev: 0000000000000000000000000000000000000000000000000000000000000000
- SHA-256 hash
- 330dc0e14959b216dd049b31217809e4db49d4f4573c0cf26dfd589fccecb21a
- Previous version’s hash
- 0000000000000000000000000000000000000000000000000000000000000000
How to check it yourself, without trusting this page
Download any version from /api/record/TRV-2026-2916 and hash its canonical field yourself, for example with shasum -a 256. You should get its content_hash. Each version’s text also ends with prev:and the hash of the version before it (version 1 points to 64 zeros). If anyone had changed a single character of any version after it was saved, those hashes wouldn’t line up.
ace