Orlando Writer Loses Book Proposal After AI Detectors Flag Human Draft

One of the four detectors tested by researchers correctly identified AI usage in 37 of 40 papers containing mixed text, yet the tools continue to generate false alarms that impact real-world hiring decisions. Ghostwriter Sally McGraw recently lost a book proposal opportunity after submitting a writing sample that multiple artificial intelligence detection tools flagged as machine-generated.

McGraw, a professional nonfiction writer with more than 20 years of experience, stated she had written the text herself without using AI assistance. She received an email from a prospective client indicating the work was flagged. McGraw said she felt angry about the dismissal but did not press the issue because she considered the job minor. The client confirmed screening the work but cited experience and fit as the primary reasons for choosing a different writer. McGraw was paid for the sample but believes the detector results cost her the full proposal.

Editors are increasingly using these tools to manage a rising volume of submissions. Neil Clarke, publisher of the science fiction magazine Clarkesworld, reported that monthly submissions rose from roughly 1,100 in 2021 to double that amount by February 2023. This increase followed the wide availability of AI agents like ChatGPT, Claude, and Gemini. Clarke estimated his own workload increased by about 30 percent due to the influx.

Research published in the International Journal for Educational Integrity in June compared the accuracy of four widely used detectors. The study analyzed 160 academic papers across four categories. Pangram emerged as the most consistently accurate tool in that specific test. However, the company CEO Max Spero noted that the detector can make mistakes when there is limited text to evaluate. Chris Callison-Burch, a professor at the University of Pennsylvania, explained that detectors often flag text that appears too predictable to the underlying language models.

Science blogger Janelle Shane encountered a 98.7 percent confidence score on a passage from her own 2019 book. She described the result as a clear false positive. Callison-Burch noted that human writing habits, such as using specific punctuation or starting sentences with certain adverbs, can trigger AI-positive results. He emphasized that detectors weigh patterns across a whole piece rather than relying on single words.

The Authors Guild stated in May that AI detectors create an environment reminiscent of witch hunts. The organization warned that false positives can damage an author's reputation and contracts. Clarke said he does not personally use detectors to screen manuscripts but believes they can help manage volume if editors verify the findings. He advised against turning suspicions into accusations.

McGraw said the fear of future accusations has altered her writing style. She now avoids certain words and constructions and occasionally omits commas to appear more human. She noted the irony of trying to make human prose look less like AI while large language models learn from that same prose. She is currently bracing for a time when a high-stakes assignment depends on the outcome of a detection scan.