Image

The new stakeholders in the music and video industry use selection algorithms to optimise the recommendations to their users. What are the main issues in the data field for the cultural sector?

Within just a few years, data has become the cornerstone of the user’s experience because it is necessary to support the user in a extensive catalogue of content. A recommendation tool is essential to facilitate its use and ensure an enjoyable experience. However, in the cultural field, a “bad” recommendation is taken very badly: it gives the impression that the service provided did not meet the expectations of the individual. All professionals in the sector will say: if you recommend a piece of music or a film that the person loathes, he or she will vehemently complain because cultural content has an identity-specific function.

Data therefore becomes a strategic resource. But which data?

The data for describing works and content, naturally: describing a film, classifying a song, etc. However, increasingly now, there is a great temptation to obtain more information about the user: his/her tastes initially, but also the setting (when travelling, heading to or from work, spending time with the family, playing sport, partying, etc.). All of that enriches the service, but this also involves more intrusion by the service into people’s private lives. And culture is a very particular sort of commodity: what we read, watch and listen to reveals a lot about ourselves, our identity and our lifestyle. Economic stakeholders must also take these elements into account in order to build a trusting relationship with their users.

The CNIL is particularly well-known for its regulator role; but is the approach to these future-oriented issues more pedagogical?

These days, the role of a regulator is multi-faceted: complaint investigations, inspections and even sanctions are a major aspect of the work, but the regulator is also there to support economic stakeholders or public authorities, to advise them and to educate these stakeholders as well as the general public. From this point of view, the future-oriented approach is an important tool, even from the regulation perspective: it demonstrates that the regulator must also take part in company debates on emerging issues, or even catalyse them. That can happen much earlier than is the case of a more legal stance and helps to build a co-regulation space with all of the stakeholders.

In what way will analysing the behaviour of online cultural services, by means of the Médiamétrie study, support the efforts of the CNIL?

If you take a future-oriented approach to the digital field, you will find out that prediction is even more impossible than elsewhere and that sudden upheavals can occur. The future-oriented approach is therefore used to attempt to plan ahead based on emerging trends and weak signals. It should also draw on creativity, intuition and imagination. However, these assumptions and scenarios must be firmly rooted in practices and usage. For us, a study like the one conducted by Médiamétrie in tandem with our more future-oriented projects allows us to look at users’ actual practices and usage, even though these practices are still very much emerging, moving and somewhat unstable.

To find out more about the work of the CNIL regarding emerging and future-oriented issues, take a look at LINC (Laboratoire d’Innovation Numérique), the CNIL’s digital innovation laboratory: http://linc.cnil.fr/

Twitter account: @LINCnil