A classic story goes as follows: Two people meet, fall in love, build a life together, and start expecting a child. The modern twist on the story is the increased difficulty of keeping the baby news quiet. Much more threatening to this goal than an unannounced family member’s visit is the omnipresence of the Internet; What online traders know or can deduce about us, and potentially share with others. You may quit social media, not look up any pregnancy or baby advice on Google, avoid online parenting forums, and even limit your baby purchases to in-store and cash-only… At best, you will feel inconvenienced, at worst, you may feel like a criminal. Still, you may find yourself facing baby-related ads and promotional offers that you were trying to avoid (Hill/Forbes 2014).

Carolina Lisboa Pinto, in her PhD research, which is part of the Consumer-ID project, intends to tackle many of the questions related to the safeguarding of digital consumers from harmful effects of online profiling. The starting point of her research is to acknowledge and address the different profiles that consumers build online and/or that are built around them, based on the data they have shared. These questions have been asked before (Helberger et al./BEUC 2021), but are still in need of more elaboration: Can we actually talk about digital consumers or do we protect consumers’ digital alter egos? Should digital consumers (or their alter egos) be protected in the same or different way online as offline consumers? What consequences should there be of the structural, architectural asymmetry online for European consumer protection frameworks?

The last question brings us back to the infamous by now benchmark of the ‘average consumer’ (Duivenvoorde 2015, Luzak 2019), which we could hardly argue was ever meant to fit consumer protection purposes, instead, it more clearly responded to the EU’s Single Market-driven objectives. It also raises questions as to the scope of paternalistic actions that policymakers should be taking, especially to protect consumers against potentially harmful consequences of online profiling (Soh 2019). 

During the kick-off workshop of the Consumer ID project in Leiden, Carolina shared her initial research hypotheses and her intention to specifically focus on the harm that profiling through digital tools could bring to consumers, further exacerbating digital asymmetry. As her research will include an empirical study, we should also anticipate having new data to examine the market effects and consumer harm of profiling tools. The results of this study are likely to provide a solid benchmark, on which to base the assessment of the effectiveness of the current European consumer protection framework, as well as against which to think of policy recommendations.

Commercial online profiling and personalisation of online information tends to aim at enhancing the persuasiveness of online communication. Communication scholars have previously considered what factors could impact persuasiveness (e.g. Cialdini’s influence scale; Kaptein et al. 2015). There are, however, many uncertainties surrounding online profiling, which currently stand in the way of effectively regulating it to improve digital fairness. One of them pertains to that digital alter ego that we may create online. If we manage not to reveal our actual preferences and characteristics in the digital world, this will undermine the whole idea of profiling user data for persuasive purposes. Another ties with the efficiency of data protection frameworks, based, amongst others, on the minimisation of data processing. Building a personalised profile requires collecting and processing a few data points. If the GDPR works as intended, minimising data collection and requiring data anonymisation, this may hinder the creation of successfully persuasive profiles.

Further questions could be raised as to the specificity and the scale of risks we intend to protect consumers from when we consider the current (and future) use of profiling tools. Carolina will likely need to delineate her research project and focus on a particular category of such risks. Whether it be risks to consumers’ freedom of contract or their data security, the risk of discrimination based on profiling (price or other contract terms discrimination) or another social injustice.

Re-balancing digital fairness requires considering which of the consumer (or data) protection tools is best suited for this task. For example, it would be relatively easy to introduce (yet another) information duty about online profiling. However, transparency may not be the solution here. First, (average) consumers are unlikely to understand the scope of and risks associated with profiling, even if profiling is disclosed to them. This would follow the current behavioural pattern of consumers tending to understate the risks of data collection (Furnell and Phippen 2012). Second, behavioural biases may work against consumers here, e.g. preference for short-term convenience, which is how they perceive the benefit of targeted, personalised advertisements (Wang et al. 2023). Third, and most importantly, there is an underlying question of profiling eroding party autonomy and distorting transactional balance, which information duties and transparency would be hard-pressed to properly mitigate.

As outlined above, online profiling poses a lot of challenges, not only of the legal kind, whilst a PhD journey presents a challenge of its own. Over the coming years, we will be closely following Carolina’s research, looking for more answers to some of these pressing questions.