The use of artificial intelligence techniques for profiling users and showing them personalized adverts or information.
Examples
- Personalized adverts appearing in social media like Facebook, or in private websites, like newspaper websites.
- Personalized services offered through the government portal
Benefits
- Having access to personalized adverts and services (customer’s perspective)
- Strengthening customer loyalty through better satisfying their needs
- Reducing the time required for searching for needed products and services
Threats related to misuse and abuse
Hidden commercial interest – We assume that everything is or will be free on the Internet. The fact is that large companies use personalized ads as a way of doing their business. Essentially, it is publicity that is based on user’s personal information, the advertisements shown are relevant to the person. In this way, private companies pay for them because it is likely that the user will access the link included in the announcement to get more information, as it is of interest for them.
Privacy policies – A main problem is what personal information is collected by companies paying for the adverts. Are users aware that such information is collected and shared with the companies? A main concern is that most users do not read entirely the privacy policies to which they sign their consent. To what extent does the user want his/her information to be accessed? The proper approach would be to allow the users to determine what can be seen from their profile and what not.
User Manipulation – If the ads are repeatedly shown to people, and additionally the products offered are of their interest, the question is how can a person’s consumption habits be manipulated through such ads? In addition, taking into account the propaganda that the user sees, whether commercial or political, the system generates a surrounding context of information for that particular user that the he or she is constantly consuming. To what extent is the user immune to and aware of the context that was purposely generated for him? Somehow all that context is influencing and training the person. It was exemplified with Cambridge Analytica, an issue of how sending personalized information could affect the choices of a country like the United States. A similar example is how the UK referendum on Brexit was manipulated, affecting the entire European Community, a population of over 500 million inhabitants.
Polarization of society – there is a risk of persons being identified by their political ideologies. Based on the choices that a person does while visiting websites, the political ideology could be determined. Later, it could happen that users only receive or can access information that supports or reinforces their own ideas. Such behavior contributes to the polarization of society.
Lack of legal framework – there is a need to enforce legal mechanisms to protect user´s rights, ensuring that ethical issues, as those discussed, are properly considered. If a legal framework is missing, how can user´s rights be protected?
Ethical Challenges
The following table summarizes ethical challenges associated with personalized adverts.
| ID | CHALLENGE | RELATED TO * |
| C1 | Lack of users´ awareness of personal data that they provide to social media and other online systems and the usage of such data for hidden commercial interest | Principle R Principle A |
| C2 | Lack of users´ awareness about the policy for sharing data adopted by private service providers | Principle R Principle A |
| C3 | Risk of manipulating user´s opinion and choices through targeted information based on intelligent analysis of their personal data | Principle R |
| C4 | Risk of polarizing society given that targeting information based on the user´s preference can produce echo chambers amplifying similar messages and can reduce the social capacity for listening to different opinions and for building consensus | Principle R |
| C5 | Lack of a legal framework ensuring the protection of users’ rights while private companies benefit from misusing personal data collected from users | Principle M |
*See Principles for more information about principles.