I am going to try to explain how the recommendation systems used by companies such as Netflix, HBO, Spotify or financial entities such as BBVA work, handling customer data, it is always understood that prior authorization from them, and applying algorithms and tools such as the machine or deep learning to manage large amounts of data and effectively advise and show users their recommendations.
The key, however, and I do not say it but some of the artificial intelligence experts who have already passed through Muy Interesante and the Blog on artificial intelligence such as Nuria Oliver, Chema Alonso, Ramón López de Mántaras or Elena Alfaro herself, maximum Responsible for BBVA’s Data and Innovation in Data Science strategy is the ethics of algorithms.
All algorithms, including those created to guide and recommend users whether to watch this or that series or movie, or whether to choose one or another financial product, are ‘manipulated and created’ by the companies themselves for lawful purposes of business to build customer loyalty and trust. And who controls the companies that create the algorithms? There I leave the question, for whoever wants to answer it.
The consulting firm PwC, for example, points out after a survey carried out with users in the field of video content, that “what is especially relevant when making a recommendation to a user is not what, but why”. And it says that consumers are smart. They know when they are being pushed to view content that is not of interest to them. Today’s audiences want to understand the “why” behind the what. They are more likely to trust what you recommend if the suggestion comes with an explanation. Why does the program have a high rating? What are the critics saying? How likely am I to enjoy this show? How did others like me value it?
37% of those surveyed by PwC say they don’t want to waste their time starting a new program they don’t like, and that personalized recommendations are too risky. Helping consumers make the decision for themselves should be the goal of companies like Netflix, and also BBVA.
And thanks to an invitation from BBVA, on March 20, I attended a talk in Madrid, in which data scientists and experts from the financial institution tried to explain how a recommendation system works in the era of artificial intelligence. High-end guests and speakers such as Francisco Vieira , president of the Superior Court of Justice of Madrid, the head of innovation at the Celler Can Roca restaurant, Heloïse Vilaseca , the art director and winner of a Goya for Snow White, Alain Bainée , and the designer of interaction Javier Cañada, explained how in their day to day decisions are made based on the data, conscious and unconscious, that they have in their hands in each of their facets and tasks.
And among them Iskra Velitchkova, the host of the event, author of the dissemination study ‘RecSys’ that was presented, and specialist in Visual Thinking at BBVA Data & Analytics, ventured to point out that “ the human mind makes two very different songs remember to the same vital moment, or that a place evokes someone very specific who has never been there. And this is a huge challenge for machines ”.
The recommendation systems, said the BBVA experts, have to go much further, and model the relationship between a person and the products they use, find measurable data that allow creating these implicit relationships to, through algorithms, give even more effective recommendations and personal in all vital aspects, including, of course, those of financial life.
And PwC analysts add that helping users navigate that vast sea of information to choose from, without manipulated algorithms, should be the ethical principle that guides companies, beyond their own commercial interests.
The consultancy also says that “as the amount of content, data, supply, proliferates faster than consumers can keep track, they need smart tools that can act as a trusted friend, giving them advice on what that they must watch so that they do not waste their time ”.
The use of smarter and more ethical algorithms in the recommendation is therefore the key so that companies, whatever the field, can achieve the loyalty and trust of their customers.
