Developed around ten years ago, mobile phone applications (Tinder, Twitter, Uber, Candy-Crush and YouTube, to name just some of the most popular ones) have become part of our daily digital lives. They provide precious services, keep us informed, entertain us and can even help us find a job or a soulmate. Designed to be easy to use, they have been met with phenomenal success and count hundreds of millions worldwide.
The hidden face of digital technology
These applications, which are presented as fun or convenient, also allow the companies that develop them to collect a significant amount of information about their users – often without their knowledge – in order to offer paid products or services for example. Their publishers (in particular tech giants Google, Apple and Facebook) make huge profits by selling and analyzing this data. Their goal is to design these applications in such a way as to capture users’ attention for as long as possible.
Director Léo Favier’s web series, "Dopamine,” available since September 2019 on the Arte TV website, explores this hidden face of digital technology. Named after the molecule responsible for pleasure, motivation, and addiction, this original program shows how these applications take advantage of certain psychological mechanisms in order to keep their users active – sometimes to the point of making them completely dependent!
The director spent many months carrying out research to develop the eight episodes of the series, which is educational without being preachy, as it highlights these mechanisms and their potential danger. This research benefited from insights from Stéphane Huot, Marc Tommasi and Aurélien Bellet, researchers in computer science and mathematics at the Inria Lille Center.
At the heart of algorithms and computers
Marc Tommasi is the head of the MAGNET (Machine Learning in Information Networks) team, comprising fifteen researchers who focus in particular on machine learning techniques, which make use of this highly valuable data. “We seek to develop algorithms for data collection and analysis that respect the privacy concerns of internet-users, for example by limiting the dissemination of such data or by making it anonymous,” explains the researcher. “We therefore incorporate an ethical dimension in developing algorithms based on machine learning techniques, and we show that it’s possible to develop applications that protect users’ privacy.”
As the head of the LOKI team, Stéphane Huot works on human-computer interaction with the ten or so members of the team. “We seek to understand how IT tools (computers, tablets, telephones etc.) are used and to improve performance or make them more user-friendly. To do so, we consider theoretical and practical questions (such as how to design a tool tailored to a need), along with technological questions, by developing an overall approach to designing interactive information systems,” says the researcher.
This highly-specialized scientific expertise caught the attention of Loïc Bouchet, the producer of “Dopamine,” who contacted the researchers in May 2018 to benefit from their insight while writing the series. “Léo Favier had carried out an impressive amount of research and had a good overall view of how applications work. He wanted our expert view to fine-tune his understanding of algorithms and machines, in particular their potential uses and technical limitations,” explains Marc Tommasi. “Over the course of our collaboration, which lasted several months, we had a series of conversations with the production team to give our opinion on the script for the series,” says Stéphane Huot. For example, we provided additional bibliographical references (articles, books etc.) and specified the origin of the concepts discussed in the commentary of certain episodes.”
Ethical questions
The series also takes a broad look at the mechanisms – most of which are well-known thanks to cognitive psychology research – that push some users to addiction. “Dopamine” also benefited from the expertise of the SCALab at the University of Lille, which specializes in this area, and whose knowledge supplemented that of the Inria researchers.
Although this type of collaboration is rather unique and atypical for researchers who are more accustomed to research carried out within their scientific community, the researchers nevertheless found the work to be fascinating. “We were very impressed by their concern for educating the general public through the series, which was both well-researched and concise. The questions posed by the production team allowed us to discuss and educate the public about the issues we face today in developing digital technology. Usually, we talk about the technical solutions to which our research contributes, rather than the ethical questions!” says Stéphane Huot.
The series was therefore an opportunity for scientists to address the societal impact of their research. “Citizens, lawyers and governments are becoming aware of the importance of personal data, as evidenced by the recent GDPR (General Data Protection Regulation). Our research shows that various technical means allow us to develop IT programs that respect people’s rights. And that alternatives to the prevailing business model, created by the American tech giants, are entirely possible,” conclude the researchers.