联合早报 2020-05-25 (Zaobao link
Translation of the article
According to OECD, the Teaching and Learning International Survey (TALIS) conducted last year reports that
teachers work an average of 46 hours a week, in which 7.5 hours are spent on marking assignments.
Lu Jia Hui (31y/o) founded Getmarked, an AI startup to help alleviate and reduce the time spent on
marking thereby creating capacity in teachers.
By providing an answer key and uploading the scanned assignments onto the Getmarked's website.
Getmarked is able to use AI to identify and automate the marking for MCQs, fill-in-the-blanks and various other worksheets and print the
markings back onto the original assignment. However, there must be a objective scoring rubric thus there are certain
limitations on subjects such as literature.
"We seek the keep the marking process as close to the original ("marking-by-hand") as possible,
where teachers have the freedom to set the scoring rubric, comment on student's answer and even have the flexibility of
choosing which questions to mark" said Mr Lu during the interview. Having spent half a year as a
relief teacher in a primary and secondary schools in Singapore, Mr Lu understands that marking forms a
big part of the teachers' workload but there is currently no available tool in the world
to aid teachers in the marking handwritten paper based assignments.
Corrections and doing revision forms an important part of learning. In the coming June,
Getmarked would be looking to work with local primary schools for their pilot.
The AI automated marking system is expected to become more "intelligent" and effective as more data is collected over time.
According to Mr Lu, while it takes Getmarked's AI marker 1-2 mins to mark every page, which might not
be faster than marking by hand, it is with externalising the time teachers spend on marking that we create
capacity in teachers for classroom innovations.