Abstract
Building on the CONSOLE framework (Tan et al., 2019) which has gained traction in crisis social media communications, the CONSOLE-D is an AI dashboard which leverages the labelled data from the earlier study for supervised learning and consistent scoring of crisis social media messages to achieve the tact and nuance required in crises. Moving beyond the data from the four aviation crises which the CONSOLE framework was derived, coupled with bridging the wisdoms of literature from medical breaking bad news protocols and crisis communication theories, the CONSOLE-D is not just a dashboard that does content analysis and scoring of messages based on the CONSOLE framework. The dashboard also draws from Large Language Models (LLMs) like OpenAI to provide evaluations and recommendations to help organisations refine their crisis social media messages so as to achieve the tact, nuance, leadership, and support required in a crisis. This research dissects the backend of CONSOLE-D (console-d.com) and provides the evidence of the rigorous human versus machine reliability tests that the dashboard underwent.
Presenters
Kevin TanAdjunct Assistant Professor, Communications and New Media, National University of Singapore, Singapore
Details
Presentation Type
Theme
KEYWORDS
Crisis communication, Artificial intelligence, Social media communication
