AI text detector Pangram has launched version 3.0, introducing a new system that classifies writing into four categories, from fully human to fully AI-generated. The company reports an impressive 99.98% accuracy rate for detecting AI-generated content and says human writing is rarely misclassified.
A University of Chicago study confirmed Pangram’s robustness against manipulation. While the update provides clearer insights into AI involvement, critics question the purpose of detection, urging a shift in education rather than policing AI use.
So…:
What situations arise for you when it’s difficult to distinguish between AI-made and human-created solutions?
Source:
Picture: