All 1 entries tagged Rebecca Mace

Using Machine Learning to offer students optional feedback on their draft essay

Using Machine Learning to offer students optional feedback on their draft essays: A joint initiative with Progressay

By Rebecca Mace (Progressay), Moktar Alqaderi (Progressay), and Dr Isabel Fischer (Reader in Information Systems at Warwick Business School)

As part of their Digital Marketing and Technologies module WBS students had this summer the opportunity to receive feedback generated by Progressay, an EdAI organisation, on their draft assignment essays. The initiative was in response to students asking for more academic writing support. The project received formal ethics approval from the university. Students who decided to participate found the recommendations received useful, e.g., ‘I found the graph that showed where the references come from very useful. It is good to see what kind of research my peers do’ and ‘I think it is relatively objective and trustable, because the different evaluations it gave were consistent with my expectations.'

We discussed the importance of ethics for this type of projects already in previous articles and blogs, e.g.:

• https://journals.sagepub.com/doi/10.1177/2043886920961782#articleCitationDownloadContainer
• https://www.wbs.ac.uk/news/five-reads-you-need-to-make-ai-ethical-and-trustworthy/
• https://blogs.warwick.ac.uk/wjett/entry/ai_ethics_for/

Rebecca Mace, from Progressay, explains here her thoughts on the ethical foundations of Progressay generally and this project in particular:

Progressay is a disrupter in the EdTech world, not only due to the technology, gamification and feedback systems it offers, but due to its deeply human desire to change things from the inside. We are educationalists, not technologists, at heart. Although we work developing EdTech and EdAI, our fundamental aim is to make things better for those who struggle to find ease of access with regards to learning. Our ethical value system is almost hardcoded into everything we do. Here is how:

• We are acutely aware of the potential for algorithmic bias and seeks to avoid this by working with schools and universities that have significant diversity in their student population. This goes a long way towards ensuring that training for the machine learning model does not reflect common problems such as race or class bias.
• We firmly believe that access to education is a human right, however, having access is so much more than having the ability to attend but feeling you can fully participate in the process. Truly understanding the teaching and learning available is fundamental to a deeper understanding of ‘access’. We facilitate this through gamified and adaptive learning activities for students.
• We focus especially upon making higher achievement a understandable process and understand what reduced transparency within the marking process can do to student aspiration. Our tool marks the essay and shows the student/lecturer in a detailed way how the grade generated was arrived at. It does this using written feedback and infographic dashboards, but also a series of targets for how to improve. Aspiration is translated into achievable reality.

• We adopt an honest and open approach that allows students, lecturers and parents access and understanding into how the system works. It presents this in understandable and easily accessible dashboards. The information it presents is designed to be immediately useable. Students and lecturers can feel informed and knowledgeable about fine grained information relating to their work. Transparency is facilitated through fairness and trust.

• Humanity, not technology, is at the core of everything we do. Education is about the quality of relationships that can be developed, fostered and maintained. We have specifically designed its entire platform to retain this educational ideal adopting an “augmented” approach, where humans are helped rather than replaced. It positively impacts upon areas such as marking workload, leaving increased time (and energy) for lecturers to focus on in depth knowledge of their students’ strengths and weaknesses. It also retains a human in the loop throughout with lecturers having the option to override the system, change feedback, offer alternative comments to their students. Furthermore, the system has been developed to promote student agency. Having deeper engagement with ones learning through transparency of grading, coupled with gamification to enhance understanding, has been shown to increase student efficacy and have positive impacts on motivation and engagement.

In short, we have deliberately and mindfully developed our platform to reflect its ethical values. At the heart is a drive to enhance social mobility by democratising access to education. Those involved in developing the platform have an in-depth understanding of educational theory, as well as years of experience teaching, lecturing, and working with students at all levels. They know what limited transparency, conscious and unconscious bias, a lack of motivation, discrimination, and reduced expectation can do to a student’s educational aspiration. It is out of personal experience and a real desire for change, that we have sought to develop a tool that speaks to these issues directly.

For more information about this project please contact:

• moktar@progressay.com
• Rebecca@progressay.com
• Isabel.fischer@wbs.ac.uk

Abigail Ball : 02 Dec 2021 09:26 |  Tags: Isabel Fischer Machin Learning Moktar Alqader Progressay Rebecca Mace Teaching |  Comments (0) |  Close comments |  Report a problem