Algorithm bias at Google appears to embrace media bias
Washington Times:
What could solve the problem is giving those searching for material the ability to direct the search for positive stories about their candidate. It should not be that hard to program that result. That would give the companies a real-time look also of how skewed media stories do not reflect voter preferences. It would be another way to check teh accuracy of polling.
Robert Epstein tried a simple experiment in the run-up to the presidential election: running searches on Google and Yahoo for political topics.Other stories have indicated the strong media bias against President Trump which I suspect the algorithms are capturing. Most of the media favored Clinton over Trump so there were likely more articles favoring her.
The results were stunning. Google searches returned twice as many pro-Hillary Clinton news articles as Yahoosearches.
Perhaps even more stunning was that men and blue-state residents saw more than double the number of pro-Clintonarticles than women and people living in red states, Mr. Epstein, of the American Institute for Behavioral Research and Technology, and Robert E. Robertson, a professor at Northeastern University, argued in a report this year.
Mr. Epstein said he is still studying what caused the bias but worries that Google’s search algorithm — a form of artificial intelligence that chooses what results a searcher is looking for — ranked pro-Clinton articles ahead of positive articles about her opponent, Donald Trump.
Those algorithms have become the modern-day Walter Cronkite or Edward R. Murrow, deciding what news reaches the eyes and ears of Americans in an increasingly Google-Facebook-Twitter media environment.
...
“The social media companies are the gatekeepers,” said Frank Foer, a writer at The Atlantic and former editor of the New Republic who has authored a book on social media’s power. “Whatever choices these companies make to elevate or bury information is very powerful and will have a big impact on what people read.”
Conservatives say they have long suspected that some of the internet giants discriminate against them and their content. They point to whistleblowers who have acknowledged they were pushed to treat conservatives differently.
The companies have denied the claims. They insist the algorithms are designed to capture the most-read news stories across the political spectrum. A computer program cannot distinguish between liberal and conservative, they say.
...
What could solve the problem is giving those searching for material the ability to direct the search for positive stories about their candidate. It should not be that hard to program that result. That would give the companies a real-time look also of how skewed media stories do not reflect voter preferences. It would be another way to check teh accuracy of polling.
Comments
Post a Comment