When given a mixture of original and general abstracts, blinded human reviewers correctly identified 68% of generated abstracts as being generated by ChatGPT, but incorrectly identified 14% of original abstracts as being generated. Cohort sizes from original abstracts (x-axis) and generated abstracts (y-axis) plotted on a logarithmic 10 scale. 3: Generated abstracts score lower than original abstracts on plagiarism detectors. They incorrectly identified 32% of generated abstracts as being real, and 14% of original abstracts as being generated. Reassuringly, there are patterns that allow it to be detected by AI output detectors, although there has been exploration of techniques to fool AI output detectors27.

Source: Forbes April 26, 2023 11:44 UTC