Morgan Meaker / Bloomberg : How AI detectors like Pangram are changing human writing: authors drop em dashes, “not X, but Y” phrasing, and words like “delve”, or add typos, to look human — As AI transforms writing and detectors like Pangram police what counts as machine-generated …
How AI detectors like Pangram are changing human writing: authors drop em dashes, "not X, but Y" phrasing, and words like "delve", or add typos, to look human (Morgan Meaker/Bloomberg)
In short
Writers are actively modifying their natural styles to bypass automated AI detectors like Pangram. To prevent their authentic work from being misidentified as artificial intelligence, creators are eliminating specific linguistic habits such as the word "delve" and em dashes. Furthermore, some authors are intentionally inserting typos into their work to convince detection algorithms of their humanity.
Key points
- Detection systems built to police artificial content are directly altering authentic huma…: Detection systems built to police artificial content are directly altering authentic human writing habits.
- Certain standard vocabulary choices like the word "delve" are being avoided due to false-…: Certain standard vocabulary choices like the word "delve" are being avoided due to false-positive fears.
- Human authors are adopting deliberate flaws simply to prove that their text was not gener…: Human authors are adopting deliberate flaws simply to prove that their text was not generated by machines.