Many sites like Google tracks user behavior around the Web and posts ads accordingly. How they track and post the ads is mystery, Researchers from Carnegie Mellon University and the International Computer Science Institute (ICSI) have said that Google’s ad-targeting algorithm discriminates between internet users. They have used an AdFisher software to simulate the browsing activity. According to this report, fake male users are shown more ads of high salaries than fake female users.
This reports also showed that the transparency tool of Google named ads settings, which allow users to view and edit the interests doesn’t revels the potentially insightful information being used for the targeted users. Due to this, that some users, who are visiting website for some substance abuse, were also shown ads of some rehab programs. This is done by tracking this information from its Ads settings tool, though this info is also not disclosed by Google.
The researchers said that the main reasons for these specific patterns are not clear, because Google ads algorithm is very complex. Google uses its data to serve the ads, but the ad buyers can also take some decisions about person’s interests and make some changes by using their own data to target specific kinds of ads. The above example however doesn’t break any specific privacy rules, though Google policy forbids targeting on the basis of health conditions.
Anupam Datta, an associate professor at Carnegie Mellon and co-author of the study, who also helped in developing the AdFisher told MIT Technology Review, “ this is a concern from a society standout point of view, because the targeted ads shown to the user some time leads them to make decisions by getting influenced by those sort of ads.”
This article was published in March this year, but this discrimination algorithm continues to make headlines now a days. This month, Google apologizes after its Photo apps tagged pictures of black people as gorillas. Researchers also found that, when you search the term CEO, Google shows only 11 percent female results. This figure is totally different from real stats as female executive making up 27 percent of US CEOs.
The researchers Carnegie Mellon University and the International Computer Science Institute (ICSI) have said that it is hard to put blame on anyone as this ad-targeting algorithm of Google is quite complex. It can be Google’s fault or the ad buyer’s, because ad buyers can also make changes according to their data and kinds of ads they want to show.