Distinguishing AI from Male/Female Dialogue

Huma Shah; K. Warwick

Distinguishing AI from Male/Female Dialogue

Huma Shah
, K. Warwick

Research output: Chapter in Book/Report/Conference proceeding › Conference proceeding

243 Downloads (Pure)

Abstract

Without knowledge of other features, can the sex of a person be determined through text-based communication alone? In the first Turing test experiment enclosing 24 human-duo set-ups embedded among machine-human pairs the interrogators erred 50% of the time in assigning the correct sex to a hidden interlocutor identified as human. In this paper we present five transcripts, in four gender blur occurred: Turing test interrogators misclassified male for female and vice versa. In the fifth, machine-human conversation artificial dialogue was branded as female teen. Did stereotypical views on male and female talk sway the judges to assign one way or another? This research is part of ongoing analysis of over 400 tests involving more than 80 human judges. Can we overcome unconscious bias and improve development of agent language?

Original language	English
Title of host publication	Proceedings of the 8th International Conference on Agents and Artificial Intelligence (ICAART2016)
Publisher	SciTePress
Pages	203-210
Volume	1
ISBN (Print)	978-989-758-172-4
Publication status	Published - 2016
Event	8th International Conference on Agents and Artificial Intelligence - Rome, Italy Duration: 24 Feb 2016 → 26 Feb 2016

Conference

Conference	8th International Conference on Agents and Artificial Intelligence
Country/Territory	Italy
City	Rome
Period	24/02/16 → 26/02/16

Bibliographical note

It will be presented at the 8th International Conference on Agents and Artificial Intelligence (ICAART2016), Rome, Italy, 24-26 February, 2016

Keywords

Computer-mediated communication
gender-blur
imitation game
indistinguishability
Loebner Prize
simultaneous comparison
Turing test.

Access to Document

ShahaicombOther version, 454 KB

Cite this

@inproceedings{c1988aee6d7e45a9b222fa7246a82041,

title = "Distinguishing AI from Male/Female Dialogue",

abstract = "Without knowledge of other features, can the sex of a person be determined through text-based communication alone? In the first Turing test experiment enclosing 24 human-duo set-ups embedded among machine-human pairs the interrogators erred 50\% of the time in assigning the correct sex to a hidden interlocutor identified as human. In this paper we present five transcripts, in four gender blur occurred: Turing test interrogators misclassified male for female and vice versa. In the fifth, machine-human conversation artificial dialogue was branded as female teen. Did stereotypical views on male and female talk sway the judges to assign one way or another? This research is part of ongoing analysis of over 400 tests involving more than 80 human judges. Can we overcome unconscious bias and improve development of agent language?",

keywords = "Computer-mediated communication, gender-blur, imitation game, indistinguishability, Loebner Prize, simultaneous comparison, Turing test.",

author = "Huma Shah and K. Warwick",

note = "It will be presented at the 8th International Conference on Agents and Artificial Intelligence (ICAART2016), Rome, Italy, 24-26 February, 2016; 8th International Conference on Agents and Artificial Intelligence ; Conference date: 24-02-2016 Through 26-02-2016",

year = "2016",

language = "English",

isbn = "978-989-758-172-4",

volume = "1",

pages = "203--210",

booktitle = "Proceedings of the 8th International Conference on Agents and Artificial Intelligence (ICAART2016)",

publisher = "SciTePress",

address = "Portugal",

}

TY - GEN

T1 - Distinguishing AI from Male/Female Dialogue

AU - Shah, Huma

AU - Warwick, K.

N1 - It will be presented at the 8th International Conference on Agents and Artificial Intelligence (ICAART2016), Rome, Italy, 24-26 February, 2016

PY - 2016

Y1 - 2016

N2 - Without knowledge of other features, can the sex of a person be determined through text-based communication alone? In the first Turing test experiment enclosing 24 human-duo set-ups embedded among machine-human pairs the interrogators erred 50% of the time in assigning the correct sex to a hidden interlocutor identified as human. In this paper we present five transcripts, in four gender blur occurred: Turing test interrogators misclassified male for female and vice versa. In the fifth, machine-human conversation artificial dialogue was branded as female teen. Did stereotypical views on male and female talk sway the judges to assign one way or another? This research is part of ongoing analysis of over 400 tests involving more than 80 human judges. Can we overcome unconscious bias and improve development of agent language?

AB - Without knowledge of other features, can the sex of a person be determined through text-based communication alone? In the first Turing test experiment enclosing 24 human-duo set-ups embedded among machine-human pairs the interrogators erred 50% of the time in assigning the correct sex to a hidden interlocutor identified as human. In this paper we present five transcripts, in four gender blur occurred: Turing test interrogators misclassified male for female and vice versa. In the fifth, machine-human conversation artificial dialogue was branded as female teen. Did stereotypical views on male and female talk sway the judges to assign one way or another? This research is part of ongoing analysis of over 400 tests involving more than 80 human judges. Can we overcome unconscious bias and improve development of agent language?

KW - Computer-mediated communication

KW - gender-blur

KW - imitation game

KW - indistinguishability

KW - Loebner Prize

KW - simultaneous comparison

KW - Turing test.

M3 - Conference proceeding

SN - 978-989-758-172-4

VL - 1

SP - 203

EP - 210

BT - Proceedings of the 8th International Conference on Agents and Artificial Intelligence (ICAART2016)

PB - SciTePress

T2 - 8th International Conference on Agents and Artificial Intelligence

Y2 - 24 February 2016 through 26 February 2016

ER -

Distinguishing AI from Male/Female Dialogue

Abstract

Conference

Bibliographical note

Keywords

Access to Document

Fingerprint

Cite this