The AI engine behind ChatGPT is a remarkable thing. It can certainly be used to create writing that, on initial perusal, looks to be the work of a human. Cue dire predictions that much of the internet could be AI creations. We’ll see— however, it is important that we spot ChatGPT’s work from the real thing, particularly for academic and scientific papers. Early efforts have been patchy, but researchers at the University of Kansas have developed an algorithm they claim is 99 percent accurate. It monitors the habit of academic writers to use certain sentence constructions and ambiguous language. Will people modify ChatGPT to fool it? Nobody wants scientific papers, even if just college essays, to be AI-generated.
–CL