Dr Hilary Davis, Swinburne University of Technology
Date and Time
Theme / Track
Presentation Format
We report on an Artificial Intelligence enabled translation service used to support older people from culturally and linguistically diverse (CALD) backgrounds living at home. The overall aim is to better report on and tailor services for clients with no/limited English, and respond to increased staff workloads arising from new regulations. The translation service is delivered via mobile phone by Care Partners at our partnering organisation, and via an audio-visually enabled standalone device called ‘CareWindow’ embedded in the client’s home. CareWindow provides audio and written translation in real time, from the client’s language to staff and back, in audio-visual format. Clients/staff can listen to and read translation almost simultaneously, replacing the inconvenience and cost of human translation services. AI-generated translation summaries are automatically uploaded into the organisation’s Client Management System, TurnPoint, where they can be reviewed and edited by staff, and enacted upon for service provision. We conducted in-office observations, follow-up interviews and short surveys with staff to ascertain impacts of the initiative. The most popular languages Italian, Greek, Arabic, Polish, reported a translation accuracy rate of 90%+ in the back-end data. Staff reported that the service saves time, and improves client-staff interaction. However, hidden behind-the-scenes staff and organisational support underpins staff-client interactions. This included home-based technology set-up, client interaction guidance, and negotiating AI-human audio-textual misinterpretations. Staff particularly valued the AI-generated key recommendations, but had to edit these for the home-care context. We outline staff strategies to navigate challenges, accurately capture client-staff interaction, and help create efficiencies for future service delivery.
Keywords
Authors