AI Writing's Biggest Giveaway Has Changed: The Economist Report Reveals New Detection Patterns

Forget em-dashes: A viral report on AI-generated writing has surprising new clues

While there are some widely regarded tells of AI writing, those stereotypes may not be as reliable as they seem. A new report by The Economist analyzed the state of AI writing in 2026, identifying the telltale AI patterns readers should look out for, plus the red herrings that don't actually point toward AI. The Economist compared its own articles to versions of the same articles generated by top AI models, including OpenAI's ChatGPT, Anthropic's Claude, Google's Gemini and XAI's Grok. Its sample also included writing from other news outlets like The New York Times and The Washington Post, as well as excerpts from popular novels published between 1950 and 2022. Altogether, it compared 55,940 sentences and 1.2 million words. What not to look for Perhaps the most infamous sign of AI writing is its overuse of em-dashes. Though they were long a favorite convention of writers, the rise of AI led many people to strike the punctuation mark from their prose, for fear of readers assuming their work wasn't human-made.

Em dash: AI writing's biggest giveaway has changed

LinkedIn has added a tool for users to report "AI slop," as a report by The Economist said common signs of machine-written text are shifting. The Economist analyzed AI writing in 2026 by comparing its articles with versions generated by OpenAI's ChatGPT, Anthropic's Claude, Google's Gemini and xAI's Grok. The sample also included writing from The New York Times, The Washington Post and excerpts from novels published between 1950 and 2022. In total, the comparison covered 55,940 sentences and 1.2 million words. The report said em dashes are no longer a reliable marker of AI-generated writing. Of the major models tested, only Claude used em dashes more often than human writers. A lack of punctuation was a stronger indicator, the report said. It found large language models used fewer commas, semicolons and parentheses than humans, and often produced long sentences while overusing the word "and." The report also found that AI-generated text was often more verbose than human writing. It said the models used rarer words, scientific language, polysyllabic terms and nominalizations more often than people did. Large language models also favored rhetorical patterns such as "it's not X, it's Y" and the rule of threes, according to the report. That contributed to long sentences with limited variation in length and blocky paragraphs with uniform sentence structure. The Economist said the patterns are changing as models evolve and train on more human writing. ChatGPT, which previously used em dashes heavily, now uses them less than any other model tested and less often than humans. The report said that, for now, the clearest signs of AI-written text are verbosity and light use of punctuation.