Is Bias Avoidable in Enterprise Search? (Video)
At Enterprise Search & Discovery 2018, co-located with KMWorld 2018, Marydee Ojala, editor in chief of Online Searcher magazine, moderated a panel discussion about the issues most relevant to today’s enterprise search and discovery environment.
The discussion featured Susan E. Feldman, president, Synthexis and Cognitive Computing Consortium, and Eric Smith, chief technologist & information security officer, FireOak Strategies, LLC.
“I don’t know anybody who is unbiased. Do you?” said Feldman. “People build algorithms, people build systems, people read search results, and their views of the world are consistent with themselves but may actually be injurious or offensive to other people, and that goes for inside the organization and outside.”
Bias itself is unavoidable, Feldman noted. “There’s a great deal of bias as soon as you deal with language, and, of course, search is a language-based technology. In terms of preventing it, how can Google do a billion searches a day and have little people sitting there reading all of the searches and the search results to make sure that they are liberal or conservative, or western-oriented or not western-oriented? I don’t think it’s physically possible, much as we might sometimes want it to happen. But the fact is that algorithms and people and matching people to information is something which is promoting bias, if you want to see it that way or not."
View the complete video.
Many speakers have made their slide decks available at www.enterprisesearchanddiscovery.com/2018/Presentations.aspx
Learn more at Enterprise Search & Discovery 2019, coming to Washington, DC, Nov. 5-7.