Google’s ad-targeting algorithms show ads for higher paying jobs to more men than women — or at least that’s what was reported in a recent study from researchers at Carnegie Mellon University and the International Computer Science Institute (ICSI).
Using custom-made software to simulate web browsing activity, the researchers found ads for jobs promising large salaries were more frequently shown to male users than female users when visiting job sites.
Anupam Datta, co-author of the study, suggests Google’s ad-targeting algorithms show signs of discrimination and a lack of transparency:
“I think our findings suggest that there are parts of the ad ecosystem where kinds of discrimination are beginning to emerge and there is a lack of transparency.”
The lack of transparency Datta is referring to stems from evidence which he suggests shows Google is tracking sensitive information about a user’s search history.
For example, during this study it was found that users visiting websites about substance abuse were later shown ads for rehab programs. Nowhere in Google’s Ads Settings tool does it disclose it’s tracking this kind of information about users.
Researchers admit it’s difficult to assign absolute blame to Google’s ad-targeting algorithms in each of these cases. Google may not be the one to blame for discrimination in ad targeting, it may be the individual advertiser’s fault for targeting the ads in such a way.
Another possibility suggested in the study is that this could be a result of unpredictable machine learning, and neither Google or its advertisers are intentionally targeting ads this way.
“…even without advertisers placing inappropriate bids, large-scale machine learning can behave in unexpected ways,” the study states.
The study concludes with researches saying they cannot claim Google is guilt of violating its policies, but instead suggest Google has “lost control” over its advertising system.
No comments:
Post a Comment