Personalized ads: Difference between revisions

Line 3:

==How it works==

Some products or services offer personalized ads via ad providers, like Google Ads. ~~In order to provide personalized ads to an~~ user, data ~~of them~~ (~~and maybe their~~ consent) ~~is required~~. Some examples of data collected to provide ~~personalized ads are:~~ GPS location, age, gender, search history, ad interaction and application interactions. ~~These gathered data allow the~~ ad providers ~~to show the user ads that better fit with the possible user's needs or preferences. Some products or services~~ allow ~~the~~ users to ~~disable or~~ delete the gathered data ~~used~~ to ~~show~~ personalized ads.

Some products or services offer personalized ads via ad providers, like Google Ads. Targeted advertising works using a combination of gathered user data (that might be collected without user's consent), training data and prediction algorithms.

Some examples of data collected to provide targeted advertising consists of GPS location, age, gender, search history, ad interaction and application interactions. In some cases, ad providers allow users to delete the gathered data or to disable personalized ads.

Once enough data is gathered, it is processed and used to train algorithms to create a profile of the user, detecting their fingerprints, behaviors, interests and dislikings. This profile is then used to predict the user and select and deploy advertisements that might be relevant for the user.

==Why it is a problem==

Line 10:

Line 14:

===Manipulation===

Personalized ads can be used as a more effective way to encourage the user to pay for a product or a service, even if they actually don't need it. It is possible the ads shown might be malicious, misleading or false ads, and the user could fall into them.

==Data poisoning==

[[wikipedia:Data_poisoning|Data poisoning]] is a practice done by some users as a form of protest or for privacy purposes, to prevent algorithms to profile the user. It consists of modifying in purpose the behavior in order to contaminate the advertising algorithms. This reduces the precision to correctly predict user's behavior, making harder to deploy relevant ads according to the actual user's profile.

An extension available for browsers named [https://adnauseam.io/ Adnauseum] works as an [[Ad block|ad-blocker]] but also as a tool to blur someone's advertising profile. It achieves this by automatically clicking on every ad to send interaction data to the algorithms.

==Examples==

@@ Line 3: / Line 3: @@
 ==How it works==
-Some products or services offer personalized ads via ad providers, like Google Ads. In order to provide personalized ads to an user, data of them (and maybe their consent) is required. Some examples of data collected to provide personalized ads are: GPS location, age, gender, search history, ad interaction and application interactions. These gathered data allow the ad providers to show the user ads that better fit with the possible user's needs or preferences. Some products or services allow the users to disable or delete the gathered data used to show personalized ads.
+Some products or services offer personalized ads via ad providers, like Google Ads. Targeted advertising works using a combination of gathered user data (that might be collected without user's consent), training data and prediction algorithms.
+Some examples of data collected to provide targeted advertising consists of GPS location, age, gender, search history, ad interaction and application interactions. In some cases, ad providers allow users to delete the gathered data or to disable personalized ads.
+Once enough data is gathered, it is processed and used to train algorithms to create a profile of the user, detecting their fingerprints, behaviors, interests and dislikings. This profile is then used to predict the user and select and deploy advertisements that might be relevant for the user.
 ==Why it is a problem==
@@ Line 10: / Line 14: @@
 ===Manipulation===
 Personalized ads can be used as a more effective way to encourage the user to pay for a product or a service, even if they actually don't need it. It is possible the ads shown might be malicious, misleading or false ads, and the user could fall into them.
+==Data poisoning==
+[[wikipedia:Data_poisoning|Data poisoning]] is a practice done by some users as a form of protest or for privacy purposes, to prevent algorithms to profile the user. It consists of modifying in purpose the behavior in order to contaminate the advertising algorithms. This reduces the precision to correctly predict user's behavior, making harder to deploy relevant ads according to the actual user's profile.
+An extension available for browsers named [https://adnauseam.io/ Adnauseum] works as an [[Ad block|ad-blocker]] but also as a tool to blur someone's advertising profile. It achieves this by automatically clicking on every ad to send interaction data to the algorithms.
 ==Examples==