Wires crossed?: On chatbots as threats to reality monitoring

Maryanne Garry; Linda A. Henkel; Jeffrey L. Foster

doi:10.1037/mac0000208

Wires crossed? On chatbots as threats to reality monitoring

Maryanne Garry^*, Linda A. Henkel, Jeffrey L. Foster

^*Corresponding author for this work

Research output: Contribution to journal › Article › peer-review

Abstract

Generative chatbots, such as ChatGPT, offer new possibilities for supporting learning and cognition among experts and novices. But they also pose new and significant risks to reality monitoring. For instance, chatbots adopt a conversational, confident, deliberate style that makes it harder for people to distinguish true from false. What is more, ChatGPT gives priority to accomplishing the task the user presents, even if that means doing so by providing misinformation. Finally, ChatGPT generates misinformation at an unprecedented scale. It threatens our individual, interpersonal, and institutional reality monitoring. When misinformation goes undetected, it threatens our trust in each other and in our institutions.

Original language	English
Pages (from-to)	485-489
Number of pages	5
Journal	Journal of Applied Research in Memory and Cognition
Volume	13
Issue number	4
DOIs	https://doi.org/10.1037/mac0000208
Publication status	Published - Dec 2024

Keywords

chatbots
cognition
large language model
misinformation
source monitoring

Access to Document

10.1037/mac0000208

Cite this

@article{c8d853e791c848b4828e7cbf7dbc1e4c,

title = "Wires crossed?: On chatbots as threats to reality monitoring",

abstract = "Generative chatbots, such as ChatGPT, offer new possibilities for supporting learning and cognition among experts and novices. But they also pose new and significant risks to reality monitoring. For instance, chatbots adopt a conversational, confident, deliberate style that makes it harder for people to distinguish true from false. What is more, ChatGPT gives priority to accomplishing the task the user presents, even if that means doing so by providing misinformation. Finally, ChatGPT generates misinformation at an unprecedented scale. It threatens our individual, interpersonal, and institutional reality monitoring. When misinformation goes undetected, it threatens our trust in each other and in our institutions.",

keywords = "chatbots, cognition, large language model, misinformation, source monitoring",

author = "Maryanne Garry and Henkel, {Linda A.} and Foster, {Jeffrey L.}",

year = "2024",

month = dec,

doi = "10.1037/mac0000208",

language = "English",

volume = "13",

pages = "485--489",

journal = "Journal of Applied Research in Memory and Cognition",

issn = "2211-3681",

publisher = "Elsevier",

number = "4",

}

TY - JOUR

T1 - Wires crossed?

T2 - On chatbots as threats to reality monitoring

AU - Garry, Maryanne

AU - Henkel, Linda A.

AU - Foster, Jeffrey L.

PY - 2024/12

Y1 - 2024/12

N2 - Generative chatbots, such as ChatGPT, offer new possibilities for supporting learning and cognition among experts and novices. But they also pose new and significant risks to reality monitoring. For instance, chatbots adopt a conversational, confident, deliberate style that makes it harder for people to distinguish true from false. What is more, ChatGPT gives priority to accomplishing the task the user presents, even if that means doing so by providing misinformation. Finally, ChatGPT generates misinformation at an unprecedented scale. It threatens our individual, interpersonal, and institutional reality monitoring. When misinformation goes undetected, it threatens our trust in each other and in our institutions.

AB - Generative chatbots, such as ChatGPT, offer new possibilities for supporting learning and cognition among experts and novices. But they also pose new and significant risks to reality monitoring. For instance, chatbots adopt a conversational, confident, deliberate style that makes it harder for people to distinguish true from false. What is more, ChatGPT gives priority to accomplishing the task the user presents, even if that means doing so by providing misinformation. Finally, ChatGPT generates misinformation at an unprecedented scale. It threatens our individual, interpersonal, and institutional reality monitoring. When misinformation goes undetected, it threatens our trust in each other and in our institutions.

KW - chatbots

KW - cognition

KW - large language model

KW - misinformation

KW - source monitoring

UR - http://www.scopus.com/inward/record.url?scp=85216115169&partnerID=8YFLogxK

U2 - 10.1037/mac0000208

DO - 10.1037/mac0000208

M3 - Article

AN - SCOPUS:85216115169

SN - 2211-3681

VL - 13

SP - 485

EP - 489

JO - Journal of Applied Research in Memory and Cognition

JF - Journal of Applied Research in Memory and Cognition

IS - 4

ER -

Wires crossed? On chatbots as threats to reality monitoring

Abstract

Keywords

Access to Document

Other files and links

Cite this