Dialogue systems are AI applications widely used in many contexts requiring user interaction. However, unconstrained interaction may lead to users communicating sensitive data. This raises concerns about how these systems handle personal data, and about their compliance with relevant laws, regulations, and ethical principles. We propose to integrate advanced natural language processing techniques in a dialogue system architecture based on computational argumentation, ensuring that user data are ethically managed and regulations are respected. A preliminary experimental evaluation of our proposal over a COVID-19 vaccine information case study shows promising results.
A Preliminary Evaluation of a Privacy-Preserving Dialogue System
Fazzinga B.;
2021-01-01
Abstract
Dialogue systems are AI applications widely used in many contexts requiring user interaction. However, unconstrained interaction may lead to users communicating sensitive data. This raises concerns about how these systems handle personal data, and about their compliance with relevant laws, regulations, and ethical principles. We propose to integrate advanced natural language processing techniques in a dialogue system architecture based on computational argumentation, ensuring that user data are ethically managed and regulations are respected. A preliminary experimental evaluation of our proposal over a COVID-19 vaccine information case study shows promising results.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
