Three scientist have made a research on the ads of Google which are from Carnegie Mellon University and the International Computer Science Institute. The findings of the research showed that Google AdSense platform has the capability of distinguishing against the women looking for employment and consumers are being targeted on the basis of the information regarding their health.
Researchers built the tool for this purpose which was named as AdFisher and 21 experiments have been done by using over 17000 user profiles. The experiments had been done to analyze the impact of various users’ characteristics which are defined by Google’s Ad settings on the ads. The ads for the positions at executive level were being served for the male accounts while for the female accounts, the job postings were related to auto parts dealer, goodwill and a generic job hunting service.
More experiments showed that there were ads by its system regarding drugs and alcohol rehabilitation centers and the accounts to which these ads were served are those whose record told that substance abused websites had been browsed by them. Similarly there were accounts which visited websites regarding physical disabilities were shown ads for accessibility products.
Google has not been accused of violation of any policy as it is not an intentional act of Google. The problem is that Google has a vast system of automated advertisements and Google has been unsuccessful to have a complete control over them.
Advertisers are allowed by Google to have their ads posted on the basis of gender so it can be said that discrimination has been permitted by Google. Promotion for the highly paid jobs are more towards men but there is also a possibility that the settings of algorithms might be done by the Google which would mean that these jobs are more relevant to men and then it is probable that Google had been taught by user behavior which made the allocation of ads in the particular way. So there is a confusion prevailing regarding the blame which is to be put either on one party or more than sources.
The models being used by Google alters with the acts of the users for example if there are only men users who have been clicking on the ads related to high income job, the algorithm will start showing these ads only to them. The algorithms learning is very hard and complicated for the humans to understand.
The experiments cleared that there no standards set by Google for the advertisers as advertisers can violate the policies easily and checking for the compliance of the ads is not being done properly expect minor things like activeness of the ad links or style issues are being checked ignoring the more sensitive issues.