Enhancing student learning with custom Artificial Intelligence

Author and presenter

• Kendal Sylvester, Applied Health and Community Support, University College  

Abstract

As the shift towards online education accelerates, innovative tools are needed to provide students with practical, real-world experiences in a virtual environment. This presentation introduces 'Irene,' a custom Artificial Intelligence (AI) designed to simulate a client for students tasked with co-designing support plans. By engaging in realistic dialogue with Irene, students can refine their ideas and practice their skills in a controlled, interactive setting.

The development of Irene involved programming a Generative Pre-trained Transformer (GPT) model to replicate the responses of a real client, using transcripts from actual interviews with Irene to ensure accuracy. This process presented challenges, particularly in ensuring the AI's responses matched the client's personality and the limitations posed by the need for a paid subscription to Chat GPT, which restricts the number of interactions students can have. Despite these challenges, Irene herself has confirmed that the AI accurately reflects how she would respond stating that the AI 'made me smile' and that it 'summed me & my life pretty good! …a lot of years of satisfaction, some adventure, some daring decisions, and some sadness, but lots & lots of joy.'

This initiative allows students to engage in meaningful conversations with 'clients' on their own schedules: offering a flexible, skills-based learning experience. Attendees will gain insights into the process of creating custom GPTs for educational purposes and the potential benefits of incorporating AI into online learning environments.