AI authorship scrutiny
- Newsrooms are formalising AI disclosure and reporting alleged misuse of human bylines on AI-written articles. - Ars Technica published an AI policy requiring clear labelling, while Gizmodo reported alleged byline misuse at a newspaper company. - The parallel rise of low-cost AI book generators is intensifying scrutiny around provenance and author credibility. (arstechnica.com) (gizmodo.com) (mashable.com)
Ars Technica published a newsroom artificial intelligence policy on April 22, while Gizmodo reported the same day that McClatchy papers were attaching human bylines to some AI-written stories unless contracts blocked it. (arstechnica.com) (gizmodo.com) Ars Technica said its reporting, analysis, and commentary are “human-authored,” and that generative AI cannot be the author, illustrator, or videographer on its editorial work. The policy says humans make every editorial decision, even when staff use AI tools for parts of the workflow. (arstechnica.com) Gizmodo reported that McClatchy executives told staff the company would use journalists’ names on AI-generated articles if contracts did not give them a right to remove their bylines. The report said the rollout affected articles generated from reporters’ original work by a company tool called a “content scaling agent.” (gizmodo.com) (thewrap.com) At some McClatchy papers, the byline label already varies by contract. TheWrap reported on April 7 that Miami Herald stories were labeled “produced using AI based on original work by” a reporter, while Sacramento Bee staffers said they would withhold their bylines from similar output. (thewrap.com) The dispute centers on a basic publishing question: whether a byline identifies the person who wrote the words, the person whose reporting supplied the material, or the editor who approved publication. Ars Technica’s policy answers that by reserving authorship for humans and requiring disclosure when AI materially contributes to published work. (arstechnica.com) That question is spreading beyond newsrooms because AI writing tools are now being sold as cheap, fast ways to produce long-form books. Mashable on April 23 promoted a Youbooks lifetime plan for $34.97, and Youbooks says its Fan Plan includes 200,000 monthly credits for writing and source uploads. (mashable.com) (youbooks.com) Youbooks markets itself as an “AI Non-fiction Book Generator” that can draft up to 300,000 words from a user’s sources, with exports to formats such as DOCX, EPUB, and Markdown described in deal listings. Those pitches put more pressure on publishers and readers to ask who actually wrote a work and how much of it was generated by software. (youbooks.com) (findarticles.com) News organizations have been writing AI rules for more than two years, with researchers at Harvard Kennedy School’s Shorenstein Center cataloging policies at 52 media organizations. The April 2026 fight is more specific: not whether AI can assist reporting, but how outlets label authorship when machines produce publishable text from human work. (journalistsresource.org) (arstechnica.com) McClatchy unions have now filed grievances over the company’s AI push, according to Yahoo and other follow-up reports this week. As more outlets publish formal rules and more AI writing tools promise publish-ready copy, the byline is turning into the line readers are being asked to trust. (yahoo.com) (gizmodo.com)
