How a Reading university study fooled its examiners
University exam answers generated by artificial intelligence (AI) could be difficult to spot, even by experienced markers, a study by the University of Reading has found.
The research saw AI-generated answers submitted to examiners on behalf of 33 fake students, with the results showing it was "very difficult" to detect. It follows concerns that advances in artificial intelligence could make it easier for students to cheat.
91福利社 Radio Berkshire's Sarah Walker speaks to Dr Peter Scarfe from the university, who worked on the project. For more on this story, check the 91福利社 News app.
Photo: FG Trade/Getty Images
Duration:
This clip is from
More clips from 91福利社 Radio Berkshire
-
![]()
The end of an era in Maidenhead
Duration: 05:23
-
![]()
What's happening with the Hexagon?
Duration: 09:48
-
![]()
Hook's Hollywood stars
Duration: 06:26
-
![]()
'No overnight solution' at Twyford
Duration: 04:26
