The ordinary inadequacy of ‘conversational AI’.
Saved in:
| Title: | The ordinary inadequacy of ‘conversational AI’. |
|---|---|
| Authors: | Albert, Saul, Stokoe, Elizabeth |
| Source: | Psychologist. Jun2026, p28-31. 4p. |
| Subjects: | Language models, Conversation analysis, Emergency management, Emergency medical services, Chatbots, Communicative competence, ChatGPT |
| Abstract: | This article examines how Large Language Models (LLMs) perform in emergency call interactions compared to human dispatchers, using conversation analysis (CA) to highlight key differences. It presents real and simulated emergency calls, including the well-known “pizza call,” where a caller covertly requests help, demonstrating that earlier LLMs failed to recognize such nuanced distress signals, while more recent versions like ChatGPT-4o show improved but imperfect understanding by relying on pattern matching from extensive training data rather than genuine interactional competence. The authors argue that LLMs lack the practical, moment-to-moment conversational skills essential for interpreting unique, context-dependent emergencies, emphasizing that despite advances, human dispatchers remain better equipped to handle the complexities of emergency communication. The article also critiques the common industry notion of “conversational AI” as overlooking the fundamental social mechanics of real conversation that LLMs cannot fully replicate. [Extracted from the article] |
| Copyright of Psychologist is the property of British Psychological Society and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.) | |
| Database: | Psychology and Behavioral Sciences Collection |
|
Full text is not displayed to guests.
Login for full access.
|
|
| Abstract: | This article examines how Large Language Models (LLMs) perform in emergency call interactions compared to human dispatchers, using conversation analysis (CA) to highlight key differences. It presents real and simulated emergency calls, including the well-known “pizza call,” where a caller covertly requests help, demonstrating that earlier LLMs failed to recognize such nuanced distress signals, while more recent versions like ChatGPT-4o show improved but imperfect understanding by relying on pattern matching from extensive training data rather than genuine interactional competence. The authors argue that LLMs lack the practical, moment-to-moment conversational skills essential for interpreting unique, context-dependent emergencies, emphasizing that despite advances, human dispatchers remain better equipped to handle the complexities of emergency communication. The article also critiques the common industry notion of “conversational AI” as overlooking the fundamental social mechanics of real conversation that LLMs cannot fully replicate. [Extracted from the article] |
|---|---|
| ISSN: | 09528229 |
