東華大學圖書館 |

Language: English

Help

回圖書館首頁

手機版館藏查詢

Back

Switch To: Labeled | MARC Mode | ISBD

Artificial Intelligence Has an Impli...

Thompson, Erica,

Linked to FindBook

Google Book

Amazon

博客來

Artificial Intelligence Has an Implicit Bias Problem /

Record Type:	Electronic resources : Monograph/item
Title/Author:	Artificial Intelligence Has an Implicit Bias Problem // Erica Thompson.
Author:	Thompson, Erica,
Description:	1 electronic resource (64 pages)
Notes:	Source: Masters Abstracts International, Volume: 81-11.
Contained By:	Masters Abstracts International81-11.
Subject:	Information technology. -
Online resource:	https://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=27963260
ISBN:	9798644902576

Artificial Intelligence Has an Implicit Bias Problem /
Thompson, Erica,

Artificial Intelligence Has an Implicit Bias Problem /Erica Thompson. - 1 electronic resource (64 pages)

Source: Masters Abstracts International, Volume: 81-11.

The algorithms underlying many artificial intelligence products have a history of implicit bias against women, people of color, and other minority groups. Research has shown AI-based systems frequently discriminate against female users with a tendency to default towards male selection, and against users of color with a tendency to default towards white selection. This implicit bias in artificial intelligence algorithms arises from implicit, and sometimes explicit, biases of the culture which created it. Additionally, the ubiquity and opacity of algorithm use mean algorithmic biases are uniquely positioned to affect protected classes of people and more deeply engrain existing cultural biases into future AIs.This paper used a comprehensive review of available literature on algorithmic bias to develop a proposed framework to identify and eliminate bias in artificial intelligence products. The Mosaic Framework calls for an intradisciplinary approach to identifying algorithmic bias and mitigating its harm. This approach mandates additional diversity within AI development and test teams; diverse, inclusive, and representative data for AI training; and independent human and AI auditors to observe, record, and review the algorithms' outcomes. In an ideal world, bias could be eliminated before it entered the algorithm; however, until that time, reducing algorithmic bias by applying an anti-bias methodology will increase the user's confidence in artificial intelligence - a necessity as it becomes progressively more interwoven in culture and society.

English

ISBN: 9798644902576Subjects--Topical Terms:

532993
Information technology.
Subjects--Index Terms:

Algorithmic bias

Artificial Intelligence Has an Implicit Bias Problem /
LDR:02922nmm a22004453i 4500 001 2435565
005 20251009122828.5
006 m o d
007 cr|nu||||||||
008 261011s2020 miu||||||m |||||||eng d
020 $a 9798644902576
035 $a (MiAaPQD)AAI27963260
035 $a AAI27963260
040 $a MiAaPQD $b eng $c MiAaPQD $e rda
100 1 $a Thompson, Erica, $e author. $3 3825898
245 1 0 $a Artificial Intelligence Has an Implicit Bias Problem / $c Erica Thompson.
264 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2020
300 $a 1 electronic resource (64 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Masters Abstracts International, Volume: 81-11.
500 $a Advisors: Darraj, Emily Committee members: Morales, Ismael.
502 $b M.S. $c Utica College $d 2020.
520 $a The algorithms underlying many artificial intelligence products have a history of implicit bias against women, people of color, and other minority groups. Research has shown AI-based systems frequently discriminate against female users with a tendency to default towards male selection, and against users of color with a tendency to default towards white selection. This implicit bias in artificial intelligence algorithms arises from implicit, and sometimes explicit, biases of the culture which created it. Additionally, the ubiquity and opacity of algorithm use mean algorithmic biases are uniquely positioned to affect protected classes of people and more deeply engrain existing cultural biases into future AIs.This paper used a comprehensive review of available literature on algorithmic bias to develop a proposed framework to identify and eliminate bias in artificial intelligence products. The Mosaic Framework calls for an intradisciplinary approach to identifying algorithmic bias and mitigating its harm. This approach mandates additional diversity within AI development and test teams; diverse, inclusive, and representative data for AI training; and independent human and AI auditors to observe, record, and review the algorithms' outcomes. In an ideal world, bias could be eliminated before it entered the algorithm; however, until that time, reducing algorithmic bias by applying an anti-bias methodology will increase the user's confidence in artificial intelligence - a necessity as it becomes progressively more interwoven in culture and society.
546 $a English
590 $a School code: 1754
650 4 $a Information technology. $3 532993
650 4 $a Gender studies. $3 2122708
650 4 $a African American studies. $3 2122686
653 $a Algorithmic bias
653 $a Artificial intelligence
653 $a COMPAS
653 $a Gender
653 $a Online advertising
653 $a Race
690 $a 0489
690 $a 0733
690 $a 0296
710 2 $a Utica College. $b Cybersecurity. $e degree granting institution. $3 3817542
720 1 $a Darraj, Emily $e degree supervisor.
773 0 $t Masters Abstracts International $g 81-11.
790 $a 1754
791 $a M.S.
792 $a 2020
856 4 0 $u https://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=27963260