This article is part of our series on trust, published on the occasion of the release of the Fondation Mines-Télécom booklet: “The new balances of trust: between algorithms and social contract.”
What does “trust” mean today? Technology is redefining the concept of trust by placing the issue of identity transparency in the forefront. But according to Armen Khatchatourov, a philosopher with Télécom École de Management and member of the IMT chair Values and policy of personal information, the question is much deeper, and this approach is problematic. It reveals how trust is reduced to mere risk management in our social interactions. This new perception is worth challenging, since it disregards the complexity of what trust truly is.
In our digital societies, the issue of trust seems closely linked to identity transparency. In order for us to trust something, must we know who is behind the screen?
Armen Khatchatourov: It seems to me that reducing trust to an issue of identity is very simplistic. This association is increasingly present in our society, and bears witness to a shift in meaning. Trust would mean being able to verify the identity of those we interact with on digital networks—or at least their legitimacy—through institutional validation. This could mean validation by a government system or via technology. Yet this offers a view of trust that is solely an assessment of the risk taken in the context of interaction, and suggests that increasing trust is, above all, a matter of reducing risks. Yet in my opinion, this approach is very simplistic in that trust is not such a uniform concept; it includes many other aspects.
What are the other aspects involved in trust?
AK: Work by sociologist Niklas Luhmann shows that there is another form of trust that is rooted in the interactions with those who are close to us. This typically means trust in the social system as a whole, or the trust we build with our friends or family. A child trusts his or her parents for reasons that are unrelated to any risk calculation. Luhmann used two different words in English—“trust” and “confidence”—which are both translated by one word in French: “confiance”. According to Luhmann, this nuance represented a reality: “trust” can be used to describe the type of assurance that, in its most extreme form, can be described as pure risk management, whereas “confidence” is more related to social interactions with those close to us, a type of attachment to society. However, things do not seem as consistent when we consider that both terms can apply to the same relationship. I would tend to use “confidence” in describing the relationship with my friends. But if I decide to create a startup with them, what I experience would be more appropriately described as “trust”. The opposite can also be true, of course, when repeated interactions lead to an attachment to a social system.
Does the idea of “trust” take precedence over the concept of “confidence”?
AK: Unfortunately, the difference between these two terms related to trust tends to be lost in our society, and there is a shift towards one standardized concept. We increasingly define trust as recommendations that are combined to form a rating on an application or service, or as a certification label. Economic theory has thematized this in the concept of information asymmetry reduction. Here we see the underlying conceptual framework and the primarily economic notion of risk it is associated with. Incidentally, this form of “trust” (as opposed to “confidence”) is based on opaque mechanisms. Today there is an algorithmic aspect that we are not aware of in the recommendations we receive. The mechanism for establishing this trust is therefore completely different from the way we learn to trust a friend.
So, is identity transparency a non-issue in the discussion on trust?
AK: Some people are reluctant to embrace pseudonymity. Yet a pseudonym is not a false identity. It is simply an identity that is separate from our civil identity, as defined by our identity card. In a sense, you have a sort of pseudonym in all traditional social relationships. When you meet someone in a bar and you develop a friendly or romantic relationship, you do not define yourself according to your civil identity, and you do not show your ID. Why should this be different for digital uses?
Aren’t there instances where it remains necessary to verify the identity of the individuals we are interacting with?
AK: Yes of course. When you buy or sell a house you go through a notary, who is a trusted third party. But this is not the issue. The real problem is that we increasingly have a natural tendency to react with an attitude of distrust. Wondering about the identity of the person offering a ride on the Blablacar carpooling website illustrates this shift: no one who is hitchhiking would ask the driver for his or her ID. What didn’t seem to pose a problem a few years ago has now become problematic. And today it is unheard of to say that transparency is not necessarily a sign of confidence, yet this is precisely the kind of issue we should be discussing.
Why should this shift be challenged?
AK: Here we need to look at the analysis, the approach at the heart of philosopher Michel Foucault’s work, of things that seemed to go without saying at a given time in history, from underlying mechanisms to representations accepted as essential components. He particularly examined the transition from one construction to another, the historical evolution. We are likewise in the midst of a new system, in which something like a “society” is attainable via social interactions. This shift in the themes of identity and trust bears witness to the changes taking place in society as a whole, and the changes in social connections. And this is not simply a matter of risk management, security, or economic efficiency.
Isn’t this identity-focused trust crisis contradictory in a context of personal data protection, which is increasingly necessary for new digital services?
AK: Yes, it is, and it’s a contradiction that illustrates the strains on the notion of identity. On the one hand, we are required to provide data to optimize services and demonstrate to other users that we are trustworthy users. On the other hand, there is an urge to protect ourselves, and even to become withdrawn. These two movements are contradictory. This is the complexity of this issue: there is no one-way, once-and-for-all trend. We are torn between, on the one side, a requirement and desire to share our personal data—desire because Facebook users enjoy sharing data on their profiles—and, on the other side, a desire and requirement to protect it—requirement because we are also driven by institutional discourse. Of course, my position is not against this institutional discourse. GDPR comes to mind here, and it is, as we speak, most welcome, as it provides a certain level of protection for personal data. However, it is important to understand the broader social trends, among which the institutional discourse represent only one element. These tensions surrounding identity inevitably impact the way we represent trust.
How does this affect trust?
AK: The chair Values and policy of personal information that I am a part of led an extensive inquiry with Médiamétrie on these issues of trust and personal data. We separately assessed users’ desires to protect their data, and their sense of powerlessness in doing so. The results show a sense of resignation among approximately 43% of those questioned. This part of the inquiry is a replication of a study carried out in 2015 in the United States by Joseph Turow and his team, in which they obtained results of a sense of resignation among 58% of respondents. This resignation results in individuals providing personal information not to gain an economic advantage, but rather because they feel it is unavoidable. These results inevitably raise the question of trust in this report. This is clearly an attitude that contradicts the assumptions some economists have made that the act of providing personal data is solely motivated by a cost-benefit balance the individual can gain from. This resignation reveals the tension that also surrounds the concept of trust. In a way, these users are neither experiencing trust nor confidence.