Kindergarten educators test an MCAST-created AI tool

A new AI database could soon help kindergarten and early-year educators prepare lessons about Malta’s marine environment.

The Supporting Early Childhood Education in Blue skills (SELBI) platform is a database powered by artificial intelligence and designed to support educators in teaching young children about the Mediterranean Sea and its importance to the Maltese islands.

SELBI works using an AI framework called Retrieval-Augmented Generation, which allows the model to find facts from the database filled with scientifically accurate, age-appropriate information focused on Malta and Maltese marine life.

Like familiar Large Language Models such as ChatGPT, teachers can ask SELBI questions, and it will give real-time answers. But unlike ChatGPT, SELBI will only use the information from the database to answer questions.

The database was developed by a team of six researchers at MCAST. All the material is tailored for educators working with children in Kindergarten 1 and 2, and Years 1 and 2.

For example, a teacher can type a request like “help me prepare a lesson for Kindergarten 1 students on Maltese fish species” and SELBI will generate a full lesson plan, including learning intentions, objectives and outcomes.

The SELBI team, from left: Francis Delicata, Kimberly Terribile, Frankie Inguanez, Shirley Ann Gauci, Heathcliff Schembri and Alan Gatt.

Idea sparked by lack of early-year resources on sea

Dr Shirley Ann Gauci, the project’s lead researcher, said the idea was sparked by a lack of early-year resources focused on the sea, despite Malta’s marine sector contributing around 15 per cent of Malta’s GDP.

“During my research, I realised that early-learner learning experiences would rarely give importance to this topic,” she said.

Gauci, who holds a PhD in early childhood education, leads a multidisciplinary team of experts in marine studies, AI and pedagogy.

Over the past year, they have developed SELBI to equip teachers with resources and training to nurture “blue skills”, a term for knowledge related to the sea and sustainability.

She said teaching children about Malta’s marine environment in an age-appropriate way helps them understand that the sea is both valuable and vulnerable.

The AI database was created by the team and the marine-related content was added by marine expert Kimberly Terribile, while education specialist Heathcliff Schembri ensured that the material was suitable for learners. Francis Delicata supports the administrative part.

Example of a SELBI chat. Photo: SELBI

AI developers Alan Gatt and Frankie Inguanez created the platform’s backend to ensure the system could respond to educators’ queries in real time.

The team first conducted a nationwide survey with over 900 educators working in the years of Kindergarten 1 to Year 2. Gauci said the first stage of research helped the team understand that many lacked confidence in teaching topics related to marine habitats, coastal care and water conservation.

“So far, the team has conducted 30 test runs to ensure the database is factually accurate, child safe and suitable for educators,” she explained.

Following the trials, four educators from a pilot school are collaborating with the team and have begun testing the platform. Over the next four months, they will provide feedback to help improve it.

“If all goes well, we plan to apply for funds and be able to provide the database to all educators, so they can use it in the classroom.”

SELBI is funded by Xjenza Malta under the Research Excellence Programme.