Study of 97 Zimbabwean authors finds zero false AI flags from Pangram
Tuhin Chakrabarty's study scanned books by 97 Zimbabwean authors published 2000-2025 and found Pangram flagged zero passages as AI-written.

A new study examining whether Pangram, an AI detection tool, shows racial bias against African authors found no evidence that it does. Tuhin Chakrabarty, a professor of computer science at the State University of New York at Stony Brook, conducted the research to investigate claims that Pangram might be prone to erroneously flagging writing by Black authors as AI-generated.
Chakrabarty scanned the books of 97 Zimbabwean authors published between 2000 and 2025, covering work from both before and after the rise of large language models. Pangram labelled zero passages of these works as AI, according to Publishers Marketplace.
If you've been nervous that AI detectors might single out writers because of where they're from, this is a genuinely good data point to have. A real study, a real sample, and a clean result: zero false flags across 97 authors spanning 25 years, pre- and post-LLM. It doesn't settle every question about AI detection, but it takes one specific fear off the table for now.
From the Coral Wire edition of Wed, Nov 18 2026 · Coral Index 69/100 (Tailwinds building)