It can be assumed that since the dawn of mandatory education, kids have tried their best to cheat their way through the many tests and examinations they face. Now – as disappointing as it may sound to those who have perfected the art of cheating – artificial intelligence will now be able to tell if you have actually written an assessment yourself, or if someone else has done it for you.
Developed by a team from the University of Copenhagen, the AI boasts a 90% accuracy when it comes to determining the legitimacy of an assessment piece.
Statistically the number of high school students cheating on major tests is growing – in Denmark at least – with educators worried about the implications of handing off important work to a third party.
Denmark’s education system – along with many other countries – have attempted to combat this by using software that looks for directly pasted lines of text. While this technology has been in place for some time now, and while it certainly would have had some effect, for the most part sentences can easily be altered slightly to bypass the software, and the method of detection isn’t that advanced – websites exist that allow you to check if a paper will be flagged for copied text or not.
The team at the University of Copenhagen claims that their technology should be able to fix this problem. The program, known as Ghostwriter, uses big data and neural networks to help it understand the various elements which make every assessment piece unique from another, and what patterns tie it to a particular author. In this case the system is trained off of thousands of previously completed exams.
With the AI in place, every time a student submits new work it is then checked against any previous work the student has submitted. Searching for patterns and similarities between the new assessment and previous submissions, the system judges how words have been used, how sentences are structured, and the length of the piece – among other things. The end result is not a ‘good’ or ‘bad’, or the name of who the student cheated off, but rather a score displayed as a percentage.
ageofrobots, , Environment, Society, American Chemical Society (ACS), Electronic Waste (E-Waste), Jinhui Li, John A. Mathews, Recycling, Urban Mining, Xianlai Zeng, 0
Electronic waste — including discarded televisions, computers and mobile phones — is one of the fastest-growing waste categories worldwide.