Integrating Audio-Based Interactions and Large Language Models into Ambient Assisted Living Environments

Ambient Assisted Living (AAL) environments have received increasing interest in the last decade. One of the aspects behind such an interest is population ageing, which introduces societal, economical and technological challenges. Nevertheless, the recent advancements in the latter fields, paired with the sophistication of Artificial Intelligence, IoT, and communication systems, leverage the provision of systems to improve individuals’ quality of life, enabling independent living for longer periods, assisting people with disabilities, and supporting caregivers and medical staff. Among the technologies being explored in AAL, audio is increasingly used alone or alongside other modalities due to its interaction capabilities. Moreover, large language models (LLMs), enable enriched human-computer interactions. In this article, we present a system that leverages audio analysis in combination with LLMs in the context of AAL aiming at providing services such as cognitive assistance, medical follow-up or entertainment. Preliminary tests show promising results to promote the adoption of LLMs in AAL environments.