Dr MattTabolism – an AI assistant that helps me help students with biochemistry

Image generated by Adobe Firefly

BCMB2001 Biochemistry and molecular biology is a second-year unit of study in the School of Life and Environmental Sciences. Students find this unit challenging and struggle to make connections between complex biological processes such as metabolism. This is apparent by the questions posted on the unit’s discussion board, usually with over a thousand threads and 250,000+ views of enquiries relating to learning content.

To address this challenge, we created an Artificial Intelligence (AI) assistant called “Dr MattTabolism”. The AI assistant was built using Cogniti, a generative AI platform developed by educators at the University of Sydney. Cogniti gives educators the ability to steer powerful large language models (LLMs) like GPT-4 with specific instructions and resources to achieve discipline-specific responses that support students’ learning and development. Our “Dr MattTabolism” Cogniti agent was provided with a system prompt to act as an expert in biochemistry and metabolic pathways. The agent was linked to lecture resources from our unit’s Canvas site and asked to engage with students in Socratic dialogue. The image below shows how simple it was to design and steer the AI, and we were able to embed it directly into Canvas for seamless access by students.

Dr MattTabolism and the simple system prompt used to steer the Cogniti agent.
Dr MattTabolism is embedded in the unit’s Canvas site for easy student access.

In the brief time that Dr MattTabolism has been implemented, over 180 of 800 students have engaged in conversations about metabolism with it. Students have embraced the integration of generative AI into their learning experience, highlighting its transformative impact. Students value its interactivity, ability to offer immediate clarification, support, and personalised teaching. The Socratic dialogue approach encourages critical thinking and has been useful for information recall and as a revision tool. As a student said,

I think this is the most innovative and creative piece of technology I have engaged with. It has been so helpful in aiding me to understand concepts … Thank you so much for taking the time to set this up and I really admire and implore your dedication to furthering the studies of us students.

This innovation has been described by students as a creative and fun study aid, enhancing engagement and helping them to fill knowledge gaps quickly. Overall, there is a sense of gratitude for the implementation of Dr MattTabolism, coupled with a hopeful anticipation for its adoption across other disciplines.

This is the most amazing thing ever invented, I love that the unit coordinators are embracing the use of AI to enhance learning instead of being afraid of it.

As educators, Cogniti allows us to set up AI assistants easily and quickly even without lots of background understanding of how generative AI works, and we feel comfortable enough to deploy this to students. Cogniti makes it so much easier for us to adopt new technologies and to help our students experience the benefits of generative AI in their learning. We are grateful for the insights that can be gathered from viewing the AI’s conversation history with students — it shows what concepts students are struggling with and even more noteworthy is that it shows us how students are engaging with AI to support their learning.

Students currently don’t get specific training or opportunities to use and engage with AI in a meaningful way, even though they will need to know how to use it effectively to keep up in a rapidly changing world. We see the Cogniti agents as generative AI with training wheels since there is structure in the responses and it’s not left entirely up to the student to come up with the right prompt to benefit from AI.

Find out more

More from Matthew Clemson

Illuminating Team Culture: How Lightboards Have Revolutionised Tutor Training

Together we teach in BCMB2X01Biochemistry and Molecular Biology, a second-year unit of...
Read More