Title
Sentiment Bias in Predictive Text Recommendations Results in Biased Writing
Abstract
Prior research has demonstrated that intelligent systems make biased decisions because they are trained on biased data. As people increasingly leverage intelligent systems to enhance their productivity and creativity, could system biases affect what people create? We demonstrate that in at least one domain (writing restaurant reviews), biased system behavior leads to biased human behavior: People presented with phrasal text entry shortcuts that were skewed positive wrote more positive reviews than they did when presented with negative-skewed shortcuts. This result contributes to the pertinent debate about the role of intelligent systems in our society.
Year
DOI
Venue
2018
10.20380/GI2018.07
Proceedings of the 44th Graphics Interface Conference
Keywords
Field
DocType
bias in intelligent systems, mobile keyboards, phrase shortcuts
Predictive text,Computer science,Human–computer interaction
Conference
ISBN
Citations 
PageRank 
978-0-9947868-3-8
0
0.34
References 
Authors
0
3
Name
Order
Citations
PageRank
Kenneth C. Arnold15610.31
Krysta Chauncey200.34
Krzysztof Z. Gajos31837127.94