Paper Type

ERF

Abstract

This study investigates the evolving meanings of trust and risk in discourse about ChatGPT on Reddit. We compare cosine similarities among words embedded about ChatGPT, trust, and risk in December 2022, close to its inauguration, January 2023, one month later, and January 2024, one year after that. Our comparisons across those periods reveal a dramatic shift in discourse patterns. Initially, “ChatGPT” was mostly associated with terms often related to new IT adoption. Over time that attention partly shifted to aspects of societal impact. There was a somewhat corresponding shift among terms related to “risk”. Importantly, terms associated with “trust” initially dealt with topics mostly mentioned in past research on trust in IT, such as privacy, risk, and security, but they too shifted towards increased attention to social issues such as vulnerability, justice, equity, and fairness, suggesting a need to consider an evolving meaning of trust and risk in the context of Generative-AI. In the next versions we shall employ more complex analyses.

Paper Number

1798

Comments

SIGHCI

Share

COinS
 
Aug 16th, 12:00 AM

The Evolving Meaning of Trust and Risk in Reddit Discourse about ChatGPT

This study investigates the evolving meanings of trust and risk in discourse about ChatGPT on Reddit. We compare cosine similarities among words embedded about ChatGPT, trust, and risk in December 2022, close to its inauguration, January 2023, one month later, and January 2024, one year after that. Our comparisons across those periods reveal a dramatic shift in discourse patterns. Initially, “ChatGPT” was mostly associated with terms often related to new IT adoption. Over time that attention partly shifted to aspects of societal impact. There was a somewhat corresponding shift among terms related to “risk”. Importantly, terms associated with “trust” initially dealt with topics mostly mentioned in past research on trust in IT, such as privacy, risk, and security, but they too shifted towards increased attention to social issues such as vulnerability, justice, equity, and fairness, suggesting a need to consider an evolving meaning of trust and risk in the context of Generative-AI. In the next versions we shall employ more complex analyses.

When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.