Nuestra NEBRIJA 33 - abril 2020

26 27 Luis Miguel Pedrero Esteban Lead researcher of the School of Communication and Arts of Nebrija University interest, novelty or relevance associated with their profile and their trail of consumption on the Web. This technology is imposed as a new search pa- radigm: it is the products that find the user inter- ested in them instead of the user searching for them explicitly. This ‘delegation’ tends to dilute the individual’s critical capacity in shaping his own menu and accepting without question the propo- sals that the algorithms decide on his behalf. For years, the applica- tion of artificial inte- lligence to journalism has been used to crea- te news in an automa- ted way, the develop- ment of which is based on the combination of structured databases. This is how reports on financial results, sports, weather forecasts and even natural disasters are now published. In the early hours of March 17, 2014, a tremor occurred in the State of Califor- nia (USA) and, despite the seismograph network sending an alert to all the media, the newsrooms were empty at that time and nobody reported it... except the Los Angeles Times. A robot at that newspaper was in charge of transforming that alert into a brief note that was immediately repor- ted by the other papers and radio and TV stations on their websites. AI technologies can generate new content by sear- ching and analyzing data, but also personalize it and adapt it to the concerns and interests of di- verse audiences. One of the tools that best mate- rializes that potential is the voice (actually, natural language processing techniques or NLP) on which our interaction with commonly-called intelligent de- vices is articulated. These include smart speakers, receivers connected to the Internet and provided with a virtual assistant with capacity for learning (deep learning) that allows recognizing and atten- ding to spoken requests and, progressively, identi- fying a pattern of knowledge on which to respond with increasing adaptation to each profile. The speed with which these systems are being in- corporated into digital life exceeds that which was once achieved by smartphones: it is calculated that 224 million are already installed worldwide (double that at the end of 2018), and in 2023 their sales will be located at 12,000 million dollars (four times those of 2018). We are moving towards a future in which voice will become the usual chan- nel of technological interaction, and as these in- terfaces are normalized in the home, a new rela- tionship will be established between people and the ecosystem of onli- ne media and services. Consulting informa- tion is one of the main demands on these speakers. In Spain, ac- cording to the Associa- tion for Media Research (AIMC) in 2019, 42% of users use them to listen to news, only be- hind listening to music (76%), asking questions in general (62%) and consult the weather forecast (59%). What horizon does this scenario pose? By using the command "Hey Google, play me the news", the virtual Google Assistant lists the most prominent headlines and events at any time, ba- sed on the user's search history, location and in- terests, but not the veracity of the sources or the presence of the attributes on which the news are professionally produced. Under the efficient appearance of this persona- lized satisfaction of our curiosity for news is the hidden risk of a biased or malicious configura- tion if the content selected by the voice interfa- ces is not guaranteed to comply with the ethical principles of journalism. Researchers and pro- fessionals together must therefore undertake the challenge of designing and standardizing verification codes that guarantee the precision, neutrality, fairness and context of the informati- ve articles supplied by smart speakers. Only in this way will it be possible to prevent and combat the growing practices of misrepresentation and manipulation, the essence of fake news, and the cause of the already proven phenomena of mass persuasion for the benefit of those who use tech- nological innovation that should be at the service of transparency to shape the opinions and, spe- cially, the decisions of the citizens. The popularization of smartphones as the prefe- rred access device to the Internet has led to an irreversible change in the patterns of audiovisual consumption, in which the initiative of each indivi- dual is already imposed as a basic criterion when choosing what, when, how and through what me- dium to satisfy their likes and interests. As a result of this habit, the logics of produc- tion, distribution and commercialization of the communication industries have been redefined, which cu- rrently respond to the demands of readers, listeners and specta- tors with creations and solutions that are separated from linear program- ming that is oriented towards a la carte selection. This is how video-on-demand platforms –Netflix, HBO, Prime Video, Disney +…– have already been standardized globally, and options for listening to music streaming services such as Spotify, or au- dio content such as Apple Podcast or Google Po- dcast, have become a daily occurrence. Hence the incorporation of new players in this industry, in many cases unconnected to the traditional com- panies in the sector, and the interest of the large technology companies -Google, Amazon, Face- book, Apple- in the increasingly intense convergence between their equipment, their operating systems, the products distribu- ted in them and the te- chniques to measure how each user enjoys them. In this context, the recommendation sys- tems have gained ground as a comfortable and effective solution for making decisions in the face of an exponential offer of information and enter- tainment. By using filtering techniques designed using artificial intelligence (AI), the recommen- dation systems select the contents that may in- terest each individual the most according to the The increasing introduction of smart speakers in homes and their ability to configure our information menu through personalization filters can skew the perception and assessment of the environment if there is no guarantee that the news selected by the machines respect the criteria of rigor and truthfulness inherent to journalism. Nebrija Research B ased on filtering techniques designed using artificial intelligence , the recommendation systems select the content that may be of most interest to each individual according to their online history R esearchers and professionals must design and standardize verification codes that guarantee the veracity of the informative articles supplied by the smart speakers Artificial intelligence for secure information