Pertanika · Universiti Putra Malaysia Press

Pertanika Proceedings

An open-access proceedings series for conferences in all research fields, turning conference output into citable, DOI-assigned publications.

e-ISSN 3083-9475 4–6 pages per paper
Proceedings paper

Conversational AI-SKY companion for Emotional Support

Karnam Akhil, Vadlana Baby and Tamma Bhuwan Vamsi

https://doi.org/10.47836/pp.1.5.014
KeywordsConversational AI, emotional well-being, NLP, reinforcement learning, sky chatbot
Paper content

Abstract

In recent times, mental illness has increased exponentially which is a threat to individual’s mental well-being. Major problems such as anxiety, depression, and loneliness affect millions of people worldwide, preventing them to seek help or express their feelings. This paper proposes a conversational artificial intelligence (AI) chatbot named SKY which human-like conversations with users, that provide emotional support to people who suffering from emotional illness. The chatbot develops a safe and non-judgmental environment for users to convey their emotions openly. The chatbot can identify signs of distress, such as sadness and hopelessness, and respond with empathy and sensitivity. By using natural language processing (NLP) and machine learning (ML), SKY generates messages based on the user’s current emotional state, offering assistance when necessary. A reinforcement learning approach was deployed to state the rewards for the Bot when it receives feedback from users or engages in conversations based on the prompts.