How do we address the latest challenges posed by profiling in an AI era?

programme

1 July (Wednesday)

4.10pm-5.10pm Profiling
  • Yves POULLET, Professor Emeritus, UNamur, Council of Europe expert
  • Wojciech WIEWIOROWSKI, European Data Protection Supervisor
  • Gabriela ZANFIR-FORTUNA, Senior Counsel, Future of Privacy Forum
  • Alessandro MANTELERO, Associate Professor, Politecnico Torino - Italy

See full programme on the main page

Visit the dedicated webpage on Data Protection Views from Strasbourg in Visio (1-3 July) and see other themes:
Session 1: How to ensure that countries that commit to Convention 108+ comply with its provisions? Why do we need a follow-up and evaluation mechanism, and which one?
Session 2: How do we address the latest challenges posed by profiling in an AI era?
Session 3: What does the right to data protection imply in an educational setting? What schools have to do, and what they should stop doing?
Session 4: Are digital identity programmes being implemented with privacy by design?
Session 5: Mirror of our souls: learning Cicero’s lessons and addressing facial recognition risks
Session 6: Political campaigns and elections: why is data protection so crucial?
Speaker's presentations

Summary

Ten years after its adoption, is the Committee of Ministers’ Recommendation CM/Rec(2010)13 on the protection of individuals with regard to automatic processing of personal data in the context of profiling still relevant?

The panel finds that there is an increasing need to revise the Recommendation, mainly because AI has substantially modified the functioning of profiling as well as the scope of the risk it can generate for individuals and the society. Recent identified practice (e.g.: Cambridge Analitica case) shows how profiling can undermine democracies. In this context, the Recommendation needs to be reviewed not only by enlarging its scope and reinforcing data subjects’ rights, but also towards a possible shift from individual protection towards protection of communities and entire sectors. Examples from the field of predictive justice are provided, where the monetisation of profiling capabilities has already generated a large market in the US around the prediction of court case outcomes. Such developments, where judges are completely uninformed of being profiled, expose the whole judiciary process to unprecedented risks. The panel recommends therefore to refocus the discussion on the original meaning of jurisprudence.

By integrating AI in its way of functioning, “modern profiling” in the US has already rendered classical profiling obsolete and there are only a few “modern” legal instruments (Fair Credit Reporting Act, California Consumer Privacy Act, ACT Relating to the use of facial recognition services enacted in Washington State) to uphold the traditional values of protecting individuals and fairness in society, whereas, as the panel underlines it, it is important that these values be guaranteed by jurisprudence. When evaluating “modern profiling”, the collective societal impact needs to be taken into account in addition to transparency, the right legal basis for the processing of personal data, the role of co-designing and independent expert committees, that also need a closer watch.

The panel concludes with the suggestion that a participatory approach centred on the human being is preferable, while vigilance over algorithm is required from all actors.

The Q&A session addresses the issues of imposing constraints on specific use, labelling, certification, options for opt-out and ethic washing.