Put people in boxes

Profiling is a technique for automating the processing of your file which consists of classifying you in a category of clients. The categories are defined by the types of data: age, gender, type of device, operating system, location, browsing history, opening or clicking on emails, etc. For a long time, profiles were created a priori by company employees based on their customer knowledge.

Today, more and more companies favor algorithmic Artificial Intelligence (AI) approaches to create customer or prospect profiles whose characteristics vary over time. AI allows us to constantly learn about our behavior, our choices, our tastes, our Likes, etc., and therefore offer us an à la carte service (we speak of recommendation, or even personalization). This is the case with Facebook or Youtube, for example.

Why would we not want to be profiled?

If you consider that the company's decision harms you or is unfair, or that the services offered to you are discriminatory, you have the right to refuse to be automatically profiled, according to article 22 of the European regulations. It all depends on the automatism of the process, of course, because some organizations will retort that if you refuse to be profiled, you will not be able to benefit from their services. This is where you can go further by exercising your right to be informed:

• has a fully automated decision been made?

• what are the logic and criteria used to make the automatic decision?

You also have the right to complain:

• challenging the decision and expressing your point of view

• by requesting the intervention of a human being who can reconsider the decision

If you accept the general conditions of use and sale (CGU / CGV) of the organization, and that these stipulate that one of the processing of your data consists of profiling, then you have to do with it. If you are against it, do not accept the terms of the T & Cs and choose another organization instead. Not necessarily easy!  

​

When you apply for a loan from your bank, it calculates the risk it takes on lending you money and not getting it back. The higher the risk, the higher your rate will be. Of course, you can object to this profiling if you consider it to be harmful to you, but there is a good chance that your bank will oppose you that it is the same method for everyone and that it works like that everywhere. . If you refuse this profiling, you will certainly not be able to take out a contract with your bank, as Article 22 of the regulations points out. This is therefore the limit of this right. On the other hand, you are entitled to ask for explanations on the profiling carried out as explained above and to involve a human in the decision-making concerning your file.

Definition of law

It is the right to follow the thread of your profiling, to oppose it and to request the intervention of a human in the automated decisions concerning you.  

To minimize the effects of profiling when you surf the web, remember to regularly delete your cookies. These small files contain the traces of your previous visits and your choices, so much interesting data for the algorithms which you will cross the road without knowing it and which will offer you content, advertisements, search results according to their calculations.

But why would anyone want to be profiled automatically?

On DARE, you are not profiled. In other words, the features and results you access are the same for any user of the platform (except for the list of companies that hold your data, of course). Said like that, we say to ourselves that it seems obvious and that it should be the case everywhere. Not sure it's that simple ...

Profiling allows us to offer us services or content automatically and adapted to our behavior. First of all, it is obvious that we have the right to refuse to be “watched” and also to refuse that our past actions condition our future possibilities. So let's say that we are entitled to refuse automated profiling on principle.  

Then you have to understand the automation of decisions as a double-edged sword. On the one hand, this is what allows us to benefit, for example, from adapted recommendations and no longer search, for long minutes, for the next video to watch. On the other hand, it is what can lock us into a restricted universe of possibilities because it only depends on what we have done before. This is the dark side of automation, we find ourselves cut off from what exists elsewhere, unable to access it. This raises real questions about the freedom of access to information, whether from media sites, social networks or web search engines.

We at Syned have recently experienced this. Two of us were working together and typed the exact same words into Google. They got different results because they were separated by a few kilometers, because they didn't have the same computer, because their past trips were different. Google is not the Web, but an access portal that references sites and allows access to them. The fact that the proposed results are different poses a real problem of access to information in terms of both freedom and equity.

Automated confinement

Recommendations from Facebook, Instagram, Netflix or Youtube are good examples (among others) of profile profiling. These services only recommend products that are related to our past choices. This is how, if you have watched a few action movies on your favorite video platform, you will be stuck in an action movie universe. We like it or not.  To get out of your own profile, you have to force the algorithm to learn something else from us. It means (being able to) watch other types of films, click on other types of publications, etc. It may take a while, but the algorithm ends up putting us in another category ... which we may want to get out of one day or another ...

​

Modern times

Looking at the number of users of major social networks and other video platforms, it is obvious that the easy side outweighs the perverse effect of automated decision-making. Maybe it takes a while to observe massive awareness? In the meantime, the risks of abuses are obvious and have already been the subject of numerous articles, documentaries and studies.  

One of the best-known drifts is the Cambridge Analytica scandal, a company that exploited Facebook user data to profile American and British voters and massively influence their vote in the presidential and Brexit elections by automatically exposing them to content corresponding to their profile.

​

“Smile”

Fortunately, not all web platforms work this way. Dare's search engine is one example of alternatives. It is also a good illustration of our values.  

Two completely different users looking for the exact same thing get exactly the same results. Our engine is designed so as not to lock the user into constraining results, that is to say a vision of the web “prepared” in advance for him.  

The engine is also designed to allow anyone to access any website in the world, without limit. For us, the web must be as it was imagined in its beginnings, a space of freedom, without borders, without distortion.

Wink

Nothing to do with profiling, but we can't resist winking at other search engines who claim to be ethical or protect our privacy. It is often with regret that we discover that they actually rely on major search engines, or that their privacy policy is so vague that they do not yet inspire us with confidence ...

In addition, you should know that a web search engine has a really significant IT cost. Nothing being truly free, particularly in the digital world, it is legitimate to wonder about the means of income of these alternative engines ... In the absence of transparency, we recommend applying a simple precautionary principle, that of not take my word for it. Let's stay vigilant (without becoming paranoid though ;-)

Useful links

• Illustration of the law by the CNIL

• The official French text of the European regulations

• https://fr.wikipedia.org/wiki/Scandale_Facebook-Cambridge_Analytica

• “La tête dans la toile”, Xavier de La Porte, C&F, ISBN 978-2-915825-62-6, 512 pages, 2016

• “Les Temps Modernes” (Modern Times), Charlie Chaplin, United Artists, 86 minutes, 1936

• “Smile”, music from the end of the film “Les Temps Modernes”, 1936, composer: Charlie Chaplin (lyrics: John Turner, Geoffrey Parsons, 1954) - https://fr.wikipedia.org/wiki/Smile_(chanson)

• https://home.cern/fr/science/computing/birth-web/short-history-web

• https://freedomhouse.org/report/freedom-net