Conversational agents (CAs) have come a long way from their first appearance in the 1960s to today's generative models. Continuous technological advancements such as statistical computing and large language models allow for an increasingly natural and effortless interaction, as well as domain-agnostic deployment opportunities. Ultimately, this evolution begs multiple questions: How have technical capabilities developed? How is the nature of work changed through humans' interaction with conversational agents? How has research framed dominant perceptions and depictions of such agents? And what is the path forward? To address these questions, we conducted a bibliometric study including over 5000 research articles on CAs. Based on a systematic analysis of keywords, topics, and author networks, we derive "five waves of CA research" that describe the past, present, and potential future of research on CAs. Our results highlight fundamental technical evolutions and theoretical paradigms in CA research. Therefore, we discuss the moderating role of big technologies, and novel technological advancements like OpenAI GPT or BLOOM NLU that mark the next frontier of CA research. We contribute to theory by laying out central research streams in CA research, and offer practical implications by highlighting the design and deployment opportunities of CAs.
%0 Journal Article
%1 ls_leimeister
%A Schöbel, Sofia
%A Schmitt, Anuschka
%A Benner, Dennis
%A Saqr, Mohammed
%A Janson, Andreas
%A Leimeister, Jan Marco
%D 2023
%I Springer Nature
%J Information Systems Frontiers
%K bibliometric_analysis chatbot chatgpt conversational_agent generative_artificial_intelligence itegpub itimpub large_language_models pub_aja pub_dbe pub_jml pub_ssc voice_assistant
%P 729–754
%R 10.1007/s10796-023-10375-9
%T Charting the Evolution and Future of Conversational Agents: A Research Agenda Along Five Waves and New Frontiers
%U http://pubs.wi-kassel.de/wp-content/uploads/2024/05/JML_926.pdf
%V 26
%X Conversational agents (CAs) have come a long way from their first appearance in the 1960s to today's generative models. Continuous technological advancements such as statistical computing and large language models allow for an increasingly natural and effortless interaction, as well as domain-agnostic deployment opportunities. Ultimately, this evolution begs multiple questions: How have technical capabilities developed? How is the nature of work changed through humans' interaction with conversational agents? How has research framed dominant perceptions and depictions of such agents? And what is the path forward? To address these questions, we conducted a bibliometric study including over 5000 research articles on CAs. Based on a systematic analysis of keywords, topics, and author networks, we derive "five waves of CA research" that describe the past, present, and potential future of research on CAs. Our results highlight fundamental technical evolutions and theoretical paradigms in CA research. Therefore, we discuss the moderating role of big technologies, and novel technological advancements like OpenAI GPT or BLOOM NLU that mark the next frontier of CA research. We contribute to theory by laying out central research streams in CA research, and offer practical implications by highlighting the design and deployment opportunities of CAs.
@article{ls_leimeister,
abstract = {Conversational agents (CAs) have come a long way from their first appearance in the 1960s to today's generative models. Continuous technological advancements such as statistical computing and large language models allow for an increasingly natural and effortless interaction, as well as domain-agnostic deployment opportunities. Ultimately, this evolution begs multiple questions: How have technical capabilities developed? How is the nature of work changed through humans' interaction with conversational agents? How has research framed dominant perceptions and depictions of such agents? And what is the path forward? To address these questions, we conducted a bibliometric study including over 5000 research articles on CAs. Based on a systematic analysis of keywords, topics, and author networks, we derive "five waves of CA research" that describe the past, present, and potential future of research on CAs. Our results highlight fundamental technical evolutions and theoretical paradigms in CA research. Therefore, we discuss the moderating role of big technologies, and novel technological advancements like OpenAI GPT or BLOOM NLU that mark the next frontier of CA research. We contribute to theory by laying out central research streams in CA research, and offer practical implications by highlighting the design and deployment opportunities of CAs.},
added-at = {2023-04-22T11:07:46.000+0200},
author = {Schöbel, Sofia and Schmitt, Anuschka and Benner, Dennis and Saqr, Mohammed and Janson, Andreas and Leimeister, Jan Marco},
biburl = {https://www.bibsonomy.org/bibtex/2be7452914eafab7cb6eb3b30040ebf8f/ls_leimeister},
doi = {10.1007/s10796-023-10375-9},
interhash = {866a16c21f57a622a792bf3723191f99},
intrahash = {be7452914eafab7cb6eb3b30040ebf8f},
issn = {1387-3326},
journal = {Information Systems Frontiers},
keywords = {bibliometric_analysis chatbot chatgpt conversational_agent generative_artificial_intelligence itegpub itimpub large_language_models pub_aja pub_dbe pub_jml pub_ssc voice_assistant},
language = {English},
pages = {729–754},
publisher = {Springer Nature},
timestamp = {2024-05-23T13:28:58.000+0200},
title = {Charting the Evolution and Future of Conversational Agents: A Research Agenda Along Five Waves and New Frontiers},
url = {http://pubs.wi-kassel.de/wp-content/uploads/2024/05/JML_926.pdf},
volume = 26,
year = 2023
}