By Carlos Affonso Souza
Director, Institute for Technology and Society (ITS), Rio de Janeiro
Every day we generate more data: our schedules, itineraries, preferences, activities and even our relationships are increasingly quantified. What then is the impact of this explosion of data – potentially available for collection and analysis – on the development of new media and on freedom of expression and the press?
To discuss this challenging issue, CIMA (Center for International Media Assistance) organized a panel at the Global Media Forum, an event sponsored by Deutsche Welle in Bonn from 13 to 15 June 2016. The participants included Sumandro Chattapadhyay, from the Centre for Internet and Society (India), Lorena Jaume-Palasi, from the European Dialogue on Internet Governance, and Carlos Affonso Souza, from the Institute for Technology and Society of Rio de Janeiro. The debate was moderated by Mark Nelson, the senior director of CIMA.
Much time has been spent discussing how the Internet of Things is transforming (and will transform) the lives of people and enable the emergence of new business. In conjunction with this discussion there has been debate about what the role of the media will be in terms of exploring the implications of the generation of such huge amounts of data. In addition to emphasizing the role of the media and the impact on freedom of speech and freedom of the press, the panel also sought to emphasize issues related to the so-called Global South. This is an additional challenge since much of the discussion about Internet of Things still focuses on the experiences of developed countries.
To better understand the dilemmas involved in the debate on data governance and the role of media, Mark Nelson recalled the conclusions of Pax Technica, a book by Phil Howard, which presents the possibilities and dangers of the Internet of Things. Citing examples from the book like applications that allow citizens to demand public services such as street repair, Nelson sought to emphasize the positive use of these new tools, connecting its implementation with the so-called data explosion and its subsequent governance.
The New Digital Tripod: The Internet of Things, Big Data, and Artificial Intelligence
Carlos Affonso said that the Internet of Things concept may seem new, but that the current digital environment already incorporate the junction of three elements that are necessary to understand the dilemmas that lie ahead. The first is the massification of connected devices that communicate with each other, which is generally referred to as synonymous with the Internet of Things (IoT). In addition to connected devices, you need to add the ability to process huge volumes of data, which is where considerations about the so-called big data come into play. Finally, connected devices generate data that are processed and their conclusions are drawn through the implementation of artificial intelligence. Thus, the Internet of Things, big data, and artificial intelligence are the three elements that contribute to the complex data explosion.
Focusing specifically on data that is generated, it is also necessary to take into account four stages that allow you to build a chain of data processing. Initially there is the collection of data and at this stage the questions that are commonly asked concern who can collect this data, and is the collection done with the consent by the proprietor. Then, once collected the data is stored, thus creating questions about the safety and even the location of these data (since different countries may have different conditions for storage, and may even require data to be stored in their own country ). A third stage concerns access to the data previously collected and stored. At this point one has to question who may have access to the data and under what conditions can the proprietor seek to know how their data is being stored and treated. Treatment, therefore, would be the fourth stage, focusing on the various ways of using this data. With whom it is shared? With advertisers, business partners? What can be done with the data collected? These are questions concerning the treatment of the data.
As an example of positive use of the data in the health field, Carlos Affonso mentioned the creation of maps that depict the evolution of dengue in the state of Rio de Janeiro, which relied on big data tools to promote better processing.
Sumandro Chattapadhyay presented a series of cases that address the issue of participation in the generation and challenges of data governance. Dealing with the issue of the quality of air we breathe, Sumandro mentioned a project that measures real-time air quality in the world, and an Indian open-source initiative that allows each person to install an air quality detector at their house.
This example allows us to understand that the establishment of the Internet of Things using connected devices can hide the fundamental layer that enables its operation, which is the data itself. Once the data is quantified, in this case data on air pollution levels, varying responses can be implemented to improve public policies related to air quality, for example.
But who would be interested in data collection? Among the most obvious answers are the private sector and governments, who are seen as competitors in terms of data collection. Sumandro mentioned the current ongoing discussion in India about creating restrictions for geolocation data sharing. The requirement that information be shared with government officials spurred opposition in some sectors.
The Internet of Things in the Global South: A Focus on Collective Projects
By bringing governments to the debate on the data explosion, the issue of sovereignty soon becomes very relevant. When asked how the so-called Global South countries are facing the challenges of the Internet of Things, Lorena said there seems to be a certain division in the way different countries take ownership of this issue. In developed countries there is a certain perception that the IoT has been established more for individual use, focusing on entertainment and work. In contrast, in developing countries IoT initiatives emphasize integration to improve infrastructure. There has been an increase in tools to improve agricultural planting and harvesting, as well as several cases of success in healthcare. She gave examples of how China has been testing these techniques in African countries. At the same time, the development of plans to connect developing countries through zero-rating/sponsored access initiatives demonstrates that the boundary between public and private can be nebulous.
With regard to the governance of data and the creation of standards, Lorena reminded us that we are often used to seeing an industry create standards, which are then adopted globally. However, in the case of IoT, industry must rely on standards that were developed in a multi-sector fashion, and not just developed for a specific company or industry. This is an opportunity for the countries of the Global South to engage in the construction of the standards that will guide the evolution of IoT, rather than simply implement ready-made standards that were created and exploited without their participation.
And how is the media is affected by this new scenario? For some time there has been discussion about how the data explosion changes the perception that media outlets can gather about their audiences. At the end of the day, understanding the audience is one of the things that every company seeks to do. But how do we deal with the current situation where the key to understanding audience behavior and reaction to content necessarily involves other actors, specifically search providers and social networks? Data control issues become even more relevant in this context.
The “Filter Bubble” and the Public Sphere in the Digital Age
With regard to better understanding the audience, Carlos Affonso recalled the concept of the “filter bubble” popularized by Eli Pariser. Social networks could (should?) be a platform for people to encounter diverse content and opinions that help to form their conception of the world. On the contrary, instead of being a window, social networks can increasingly be a mirror that reflects exactly what the person already thinks. The customization of the online experience necessarily involves the collection and processing of data.
How does Netflix recommend a movie based on your preferences? From an analysis of large volumes of data that show similar interests. With the expansion of IoT it is worth remembering that even devices most would not associate with data collection, will now start actively collecting and transferring data. For example, Samsung issues a warning about its smart TVs noting t the risks of talking about personal matters in front of the TV, since the voice command function means that sound is constantly captured and processed. And we are not even talking about the case of dolls like Hello Barbie.
The risk with the addition of the tripod “IoT-Big Data-Artificial Intelligence” is the creation of a wall around individuals, which crystallizes notion about an individual’s taste and worldview that not only reinforces preconceptions, but also avoids contact with diversity.
Sumandro pointed out that in the current media environment companies stand to lose control of the data that would be relevant to audience analysis. On the other hand, this scenario provides an opportunity for the topic of data collection to become a theme for the media to cover. This is not data-driven journalism, but rather an an attempt to understand how stories that deal with the data collection issues might lead to stories that are really more important to people. What does having your data collected mean? What is the impact of data processing? What data is a new, trendy device collecting?
These questions illustrate a new challenge for the press, which should see the data explosion not just in terms of enhanced audience analysis, but also as a way to promote journalism that deals with human rights. Exploring stories on the impact of big data and IoT ultimately enlightens the public about the advantages and risks involved. The case of US big-box store Target serves as a good example. Target famously used big data analysis to predict which customers would soon become pregnant in order to target advertising. The coverage of these types of big data implementations serves as a way to understand the impact of technology in everyday life.
While on the subject of diversity and control of data, Mark Nelson questioned how the filter bubble may reduce the way in which common knowledge is distributed within the network. Freedom of the press is reflected in the diversity of opinions and comments that can be accessed by individuals.
But is the “filter bubble ” really to blame for this scenario? Lorena wondered if content filtering was not something that naturally follows the evolution of the media. In the past we shared favorite radio stations, television channels, and we chose which newspapers to read. So do we have evidence that the Internet filters content on a larger scale than the ways that we were previously filtering content?
Social Network as News Platforms?: Facebook’s Retreat from the News
In the week following the completion of the panel, Facebook announced a change in the way content posted on the social network appear in the news feed of its members.In a public statement, the company clarified that its algorithm would favor personal content, such as photos and videos of friends and relatives, at the expense of simple news sharing.
The announcement ended up taking many people by surprise since in recent years the company had been encouraging the creation of original content by media outlets on the platform, as through Instant Articles.
According to Cass Sunstein, the amendment proposed by Facebook could have been motivated by three factors: (i) as a way to combat the criticism suffered recently that the platform would be favoring certain political content and restricting the scope of conservative media outlets; (ii) to increase the number of clicks, since it would be focusing on content that, in principle, would be of greatest interest to the user; and (iii) to avoid platform transformation into a space for simple sharing news (instead of promoting the creation of personal and original content).
The Facebook announcement relates well to the issues addressed in the panel. At one point during the panel, Mark Nelson recalled the case of a journalist from Senegal who had successfully created a Facebook page for his news site, through which he was able to generate a good number of followers. However, at some point the journalist realized that readership was decreasing as people were seemingly reading his publications with less frequency. After investigating the cause of this decline, he determined that the social network had started favoring video content, which thus negatively affected the written articles he was sharing.
The example is interesting in terms of thinking about the role social networks play in communicating the diverse content, and especially with regard to their relationship with the press. Understanding the rules by which social network algorithms privilege or restrict the scope of an article is increasingly important to the online presence of media outlets. As Rafael Coimbra suggests, Search Engine Optimization (SEO) techniques that enhance the placement of content in search engines will be quickly popularized on social networks. It is now incumbent upon professionals to understand the Facebook algorithm in order to promote content at any given time.
It is precisely this scenario that raises questions about how IoT and big data will impact the way journalistic content reaches audiences. As certain social media networks and content platforms have a global impact, understanding how local media reacts and appropriates the data explosion is an increasingly interesting issue, especially in countries of the so-called Global South.
For those interested in delving into the debate, the audio of the event can be found here.
Carlos Affonso Souza is the Director at the Institute for Internet and Society (ITS) in Rio de Janeiro.
This post originally appeared in Portuguese on July 17, 2016 on the site of the Observatorio da Internet no Brasil. This text is licensed via a Creative Commons Attribution 4.o International license. The test was translated into English by Daniel O’Maley.