What is the future of personalization?
In the close future the science and practice of product recommendations has to cope different issues:
-When serving sets of recommendations, how to balance the diversity of these, how to handle unexpectedness, and other characteristics of these sets.
-Context-sensitive recommendations — more appropriacy needed to recognize and address the context in which a recommendation is called or served.
-Modelling and communicating efficiently the uncertainty in recommendations is also a field that will get data scientists work.
-A very challenging lesson will be how to balance recommendations that:
1. seem unfamiliar therefore less trusted by the customer, but they perform more usefully, but less trusted
2. against those that look familiar and invoke trust in the user.
The possible danger appearing here is the uncontrollable confluence of big data arriving from different sources however limited by several user data protection regulations cannot really be evaded. Many stores will grab the opportunity to operate with the user data gained by purchases in other stores or distributed by semi-legal and illegal channels. The big data created now and will be created in the future will make privacy and data security more important. Privacy concerns will be an enormous drag for big data and personalization industry - if the industry fails to find a solution, a long list of technologies can lose its edge and petrify hence the whole evolution can come to a standstill or even turn back. Countries such as China where user privacy protection is underdeveloped therefore big data and AI evolution sees less hurdles will bring plenties of challenges to the industry, as algorithms will not stop at the borders, will be more or less illegally used which induces more initiatives to overregulate.