Enhancing the Trust of SLPs Towards AI-based Speech Therapy Tools

Abstract

Artificial intelligence (AI) presents transformative potential for speech-language pathologists (SLPs) from early diagnostics to predictive modeling for augmentative communication and autonomous therapy; however, clinical adoption remains constrained by significant trust deficit regarding technological precision and clinical efficacy. This research addresses the socio-technical divide by introducing a human-factors-centric framework for the systematic design and evaluation of AI-driven speech therapy tools. Synthesized from the Digital Health Scorecard, the proposed framework delineates four fundamental pillars of validation: technical performance, clinical efficacy, human-centered usability, and cost-benefit transparency. The paper identifies actionable technical strategies, such as the integration of Explainable AI (XAI) and the utilization of geographically and demographically diverse training datasets to enhance predictive reliability and mitigate bias. The framework emphasizes rigorous alignment with Evidence-Based Practice (EBP) to ensure that digital interventions remain grounded in peer-reviewed clinical standards. This framework also provides a rigorous methodology for developers to align AI innovation with SLPs' professional values, facilitating a more effective integration of technology into human-centered clinical environments.

Keywords: Artificial Intelligence, Human-computer Interaction, Digital Health, Speech Therapy, Trust

DOI: 10.54941/ahfe1007988