Who wrote this? How smart replies impact language and agency in the workplace

Kilian, Wenker,

doi:10.1016/j.teler.2023.100062

Cited by 4 publications

(2 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Research by Mieczkowski and Hancock [2022] on the agency, expertise, and roles of AI systems in communication indicates that perceptions of AI's agency and expertise significantly affect AI-mediated interactions. Wenker [2023] investigates how AI-generated smart replies impact language and agency in the workplace, indicating changes in communication dynamics. These examples indicate that the nature of communication is evolving as AI systems are increasingly taking part in decision-making and content creation.…”

Section: Ai and Science Communicationmentioning

confidence: 99%

Navigating the AI era: university communication strategies and perspectives on generative AI tools

Henke

2024

Journal of Science Communication

View full text Add to dashboard Cite

This study conducts a pioneering empirical analysis of generative AI tools, such as ChatGPT, in the context of university communication across German universities. It explores the adoption rates, identifies the primary challenges, and assesses the potential of these technologies, integrating several theoretical concepts. The findings reveal a widespread use of AI for translation and language correction, with broader applications gradually emerging. Adoption rates vary significantly between private and public universities, largely due to concerns over technical issues, data protection, and AI usability. The results underscore the need for enhanced training and AI policies that support effective integration and use.

show abstract

Section: Ai and Science Communicationmentioning

confidence: 99%

Navigating the AI era: university communication strategies and perspectives on generative AI tools

Henke

2024

Journal of Science Communication

View full text Add to dashboard Cite

show abstract

“…Arnold et al (2018) showed that exposure to positive AI suggestions subsequently led to more positive human writing. More recently, experimental work showed how the presence of Smart Reply recommendations changes the speed and positivity of the replies people send and some perceptions of the communication partners and user agency (Mieczkowski et al 2021;Hohenstein et al 2023;Wenker 2023). However, bias in AI-MC has not yet been examined in large-scale and naturalistic settings that include both human and machine-generated responses.…”

Section: Bias In Ai-mcmentioning

confidence: 99%

EnronSR: A Benchmark for Evaluating AI-Generated Email Replies

Shay,

Davidson,

Grinberg

2024

ICWSM

View full text Add to dashboard Cite

Human-to-human communication is no longer just mediated by computers, it is increasingly generated by them, including on popular communication platforms such as Gmail, Facebook Messenger, Linkedin, and others. Yet, little is known about the differences between human- and machine-generated responses in complex social settings. Here, we present EnronSR, a novel benchmark dataset that is based on the Enron email corpus and contains both naturally occurring human- and AI-generated email replies for the same set of messages. This resource enables the benchmarking of novel language-generation models in a public and reproducible manner, and facilitates a comparison against the strong, production-level baseline of Google Smart Reply used by millions of people. Moreover, we show that when language models produce responses they could align more closely with human replies in terms of when responses should be offered, their length, sentiment, and semantic meaning. We further demonstrate the utility of this benchmark in a case study of GPT-3, showing significantly better alignment with human responses than Smart Reply, albeit providing no guarantees for quality or safety.

show abstract