This AI-powered marking tool will help lecturers reclaim up to 73% of their time
The first time Mohammed Alhamed was assigned to help lecturers with marking as a graduate teaching assistant at the University of Glasgow, he was welcomed by a large pile of plastic boxes containing students’ answer sheets.
“It was as tall as me,” he says. “I sat on the chair, looking at that pile and feeling rather frustrated at the thought of having to go through 400+ answer sheets to give the same feedback again and again and again.”
A PhD student in Computing Science, Alhamed knew there had to be a way to avoid such repetitive work while still guaranteeing quality in marking.
That’s where the idea for Awardee came from — an AI-powered marking tool that uses natural language processing and deep learning neural networks to help lecturers reclaim up to 73% of their time while ensuring high-quality feedback for students.
The main challenge Alhamed encountered when developing Awardee was technical: developing a model that would take into account the fact marking is far from an objective task.
“I’m not talking about true/false or multiple choice exams,” Alhamed says. “I’m talking about the subjective act of giving feedback, which has no true or false answer.”
This meant the AI model underpinning Awardee could not be trained on purely factual assignments, but it had to be developed to imitate markers’ behaviour and acquire their feedback knowledge, factoring in their subjective approach.
“What you feed is what you get, we just save you time,” he says. “If you have two lecturers marking a business essay, our tool will mark it in different ways, learning from Dave’s marking behaviour or Sarah’s marking behaviour, to reflect the fact that if you give the same essay to two different people, you’ll end up with a slight differentiation in feedback.”
Over time, individual models could be matched to allow markers to expand their feedback knowledge for the same test, paper or essay, encouraging lecturers to also learn from each other.
Alhamed joined Conception X in 2021 to figure out how to turn his research into a viable product and reach the right audience.
“During your PhD, you work in a fantasy world where everything seems attainable. Conception X opens a window onto a different world, the real world,” he says. “Before joining the programme, I was reading about entrepreneurship to understand how to make that transition; Conception X taught me how to start acting on it.”
Since being shortlisted as a finalist at the Cohort IV Demo Day, Alhamed has recruited a CTO and business developer, obtained a UK Global Talent visa that allows him to incorporate Awardee in Scotland, received office space at the University of Glasgow with the support of Student Enterprise, and is getting ready to release a beta version of the marking tool this month — soon available to test by invitation only.
The plan is to make the product available in various languages, starting with Chinese and Spanish, by 2023, build a community around it and eventually turn Awardee into a global edtech platform facilitating decentralised education (DeEd) degrees for students to acquire highly personalised transcripts.