Ethical issues of online targeted advertisements

“ Facebook removes 5000 ad targeting terms, from ‘ Passover’ to ‘ Buddhism’” an article published by Garrett Sloane on Ad Age, a leading new sources for marketing and media. Facebook, the largest social networking company leverages information it gathers on its users like your favorite pages, posts you’ve liked or events you attended. With this information, Facebook can generate classifications and send you advertisements based on your activity or personality. This is a form of targeted advertisement, these shown advertisements are curated by computer systems. The article updated the community on efforts made by Facebook to prevent bias against religion, race and cultures. A year ago, ProPublica has depicted a case where they were about to target “ Jew hater”. As a result, Facebook has looked into removing opportunities for advertisers to create ads with malicious intent and targeting people who are oppose their agenda. We will explore the ethicality of online targeted advertisement.

In the 2000s, “ many companies intensified the adoption of intrusive techniques for gathering information of users mostly without their consent” Around this time, advertisers were allowed to select who they wanted to advertise based on demographic information: location and age. Today on Facebook, you can select a host of things location, age, gender, interests, and more. In 2013, research in the understanding how these ad systems work cited that 65% of ad categories received by users are behaviorally targeted. It can be quite concerning when if we think about the information that is available on ourselves. Ad age is worried that the combinations of AI in advertisements could “ elevate the art of persuasion to the science of behavior control.” There are several other related to targeted ad issues, like the pervasiveness of the systems which consistently updates your interest profiling every 10 webpages or the persistence of a particular ad constantly shown and bombard by the same ad. Even though this is deeply intertwined “ is our data safe” is a related issue. We have seen with the Cambridge Analytica and Facebook mess, that information has been accessed to build psychological profiles. These psychological profiles were sold and used in a political campaign.

With all this information you can build behavioral patterns which plays into psychographics. Psychographics can be defined as a qualitative methodology used to describe consumers on psychological attributes; defined by psychological profile, going farther than Demographics which is generality the Location and Age group. We can say a majority of the information gather on us for the purposes of online advertisements today falls until psychographics. Ad distributors like Google and Facebook will utilize this information to deliver targeted ads. We have to define a few things first. Targeted ads, we are referring to called Interest-based advertising or online behavioral advertising is what it is known in the field of marketing and from a computer science standpoint content-targeted advertising, they are all synonymous. We will be defining systems as the computer processes as a whole from all the companies which generally collects user information, ingests user information, performs linking and producing recommended ads, the context of the text should suffice in determining which process.

There are a few stakeholders related to this ethical issue: technology companies whose main source of revenue is advertisements, online users and advertisers (small/local business and corporations). Firstly, Technology Companies I will use a couple of the biggest technology companies whose revenue is primarily from advertisements: Google and Facebook. For clarification, these companies are still considered technology companies because their services utilizes technologies despite their revenue is from advertisements; roughly 28 billion out of 32. 5 billion from advertising likewise with Facebook 13 billion out of 13. 2 billion. Online users, this includes everyone who uses online services for search, reading, shopping, etc. This group is largest of the others in population; so naturally they would be affected the most by the targeted advertisements. Advertisers, I am generalizing the advertisers into 2 groups and understand there are many complexities. We have many small businesses, local businesses and startups who rely on Google and Facebook services to drive growth. They have yet established ways to grow organically or maybe limited in their regions or growth so they rely on online adverts. I won’t expand too much on corporation because won’t be affected as much since they already have a network or audience. They also have the resources to launch national advertising campaigns.

A risk cost benefit analysis of the online advertisements would in order to understand the situation and how proposed solutions would address all these factors. There is a risk that this information being exposed or used in malicious ways, we’ve seen this information influence political agendas. These companies, Google, Facebook have already invested and developed these advertisements systems, there is a large financial and time cost already associated with it. We would Some benefits of targeted ads are: improving the user’s experience through more relevant ads, utilizing contextual information possibly to help users discover something new, leveraging Google’s and Facebook’s vast network small business / local business/ startup can reach people quicker.

Today, systems that gain information for content-targeted advertising gather more than just demographics, it gathers all the data you produce, which helps them build a roadmap of you. As a company or researcher you want to gather as much information you can and figure out what you can do with it later. Brief explanation of machine learning, why having a lot of data is important to build a model. In the feature extraction & selection is subfield of machine learning and statistics which is trying to figure out what information is important. The concept of a no free lunch theorem, there isn’t one algorithm to maximize everything; the most useful models require human-design, which could jeopardize our information.

Back in 2010, a solution was proposed by FTC to give the options to users to opt out of targeted advertisements, similar to do-not-call this number. The information is still collected, but the option to opt out would give users decisional privacy. The Digital Advertising Alliance (DAA), has created a tool that does the same thing to opt-out of targeted advertisements which allows users to choose if participating company can “ customize ads” on our browsers. DAA is not government affiliated, it is an organization that delivers self-regulatory solutions to help guide companies on different types of advertising. There are no fines, sanctions or punishment if the company in questions does something wrong, it is monitored by the Council for Better Business Bureau and Association of National Advertisers. DAA’s principles fill a gap where we did not have any community standard and guidelines on online behavioral advertisements.

Despite having standards and guidelines for companies on how to deliver ads, we want to be able to control our data. The textbook Ethics and Technology calls this informational privacy which is the control over the flow of one’s personal information. Google has a significant investment in development a robust advertisement policy and control. Google allows users to control their information and see what information Google has on them. Its settings complete with informational material about how Google uses your data and delivers ads, it even includes access to delete all information Google has on you. The only thing left out is how our information is being kept and what type of practices is used to ensure its safety in storage. A solution where we give everyone informational privacy at first would seemingly put a dent in the companies’ information warehouse.

Certainly, there are ways to inform oneself about what sites are collecting data on you. One of the ways is to use a plugin like to Collusion “ visualizing the entities that track your behavior across the web.” There are ways avoid to targeted advertisements all together like using an anonymized browser or adBlock.

Applying a rule utilitarian view would suggest promoting informational privacy and making the majority, online users happy. In this case, we do not stop the use/development of targeted ads and grant users control of their information. We may also apply an act utilitarian view because companies are also seen entities, if the company flourishes it maximizes. When the company maximizes it can in turn help others. Interestingly, in this context we could apply something akin to the ends justifies the means fallacy to the company(s) to act utilitarian, the common perception that the “ Ads are what make free products “ free””. If we were applying deontology with Google’s current privacy practices and a set of rules limiting the company to only generate revenue to maintain the services and development. The scene would look exactly like if it was out of an utopian society.

All in all, online advertising is not going away where soon. It has many forms, but targeted advertising is an ugly one. The language and verbage used by many people in the marketing and computer field feels indicative of advertisement is a necessary evil and try to pawn it off as it is improved for the user. Instead of a clean slate or holistic view of the targeted ads, people in this field are trying to grasp for our informational privacy while trying not to poke the big bear companies.