AccueilProfile, Predict and Prevent

Profile, Predict and Prevent

Data-driven policies, markets and societies

*  *  *

Publié le mardi 21 avril 2015 par Elsa Zotian

Résumé

Algorithms are increasingly used, both by States,market actors and citizens, for the purpose of profiling. Through big data analysis and inference techniques, an attempt is made to better understand, predict and, in certain cases, prevent citizen behaviour. Data analysis techniques are deployed in many sectors of society, from cyber-security and police investigations to judicial decision-making, from product customization and personalisation to marketing strategies and targeted advertising, from self-monitoring to lifestyle improvement. For this conference, we invite researchers, experts and practitioners from different backgrounds to reflect upon the legal, ethical and social implications of data-driven policies, market transactions and quantified-self techniques. We welcome empirical, theoretical and philosophical contributions regarding profiling, prediction and prevention.

Annonce

This international conference will take place at Université Paris II, on October 30-31, 2015

Argument

Algorithms are increasingly used, both by states,market actors and citizens, for the purpose of profiling. Through big data analysis and inference techniques, an attempt is made to better understand, predict and, in certain cases, prevent citizen behaviour. Data analysis techniques are deployed in many sectors of society, from cyber-security and police investigations to judicial decision-making, from product customization and personalisation to marketing strategies and targeted advertising, from self-monitoring to lifestyle improvement.

Surveillance systems allow governments to categorise and classify citizens, thereby identifying groups of so-called “risk citizens”, who are more likely to cause harm to themselves or their fellow citizens than the average citizen. The Snowden revelations have shown how the U.S. National Security Agency relies on data mining in order to preserve national security. The American Terrorism Information Awareness Program gathers and analyses data on foreign terrorists in order to preempt and defeat terrorists acts. Australian border politics aim at identifying potential visa breechers well before they reach the physical borders of the country. The Dutch police analyse mobility data of known drug traffickers in order to enhance the chance of arresting them on their way.

Markets actors are more and more interested in profiling their customers. For instance, companies such as Google and Amazon are well-known for collecting personal data and relying on sophisticated data mining techniques to increase their revenues through targeted advertising. Facebook and Twitter profile their users in order to determine which information is most relevant to their users and should therefore be displayed into their news feed. Bank and credit cards companies often make use of data mining techniques to determine their clients’ credit score before issuing a loan.

Citizens too are eager to monitor themselves. The Quantified-Self community is steadily growing, with more and more people using an array of sensors and devices to collect data about themselves and their environment. These practices are becoming especially popular in the field of personal health and chronic illness management, where big data analysis can help patients come up with a potential solution to their diseases.

Data mining and profiling allows governments and market actors to make predictions about citizens’ behaviours and whereabouts, whether these data have been provided voluntarily or not. But how do the algorithms that are used to classify citizens actually work? On what basis does someone fall into a certain category, such as “potential burglar” or “single mom”? What kind of actions can citizens undertake to have these labels be rectified or removed? What can governments do to ensure the lawfulness and fairness of these systems?

For this conference, we invite researchers, experts and practitioners from different backgrounds to reflect upon the legal, ethical and social implications of data-driven policies, market transactions and quantified-self techniques. We welcome empirical, theoretical and philosophical contributions regarding profiling, prediction and prevention.

Topics that might be considered include:

  • State of the Art: To what extent are technological systems of data-driven prediction and prevention currently in use and not merely science-fiction?

  • Implications: What are the legal implications of prediction and prevention in government and society? What are its intended and unintended consequences? What are the risks of excessive personalisation and customisation? Can it lead to a situation of preemption or exclusion from access to alternative content (e.g. filter bubble)? Can identities be (re)-construed through the analysis of anonymized data or users profiles?

  • Regulation: To what extent do technological systems behave autonomously? How much control do we have over those mechanisms? How predictable are the algorithms that predict human behaviour? How can we evaluate the legitimacy of these algorithms? Is there a need for transparency and if so, to what extent the outcome of these algorithms be manipulated? How can we evaluate the success of preventive measures, given that there is no basis to compare with the situation where no intervention had been taken?

  • Public-Private Partnership: Are government and corporate profiling complementary, supplementary or incompatible? To what extent do they rely on each other for achieving their own goals? Does a clear distinction between public and private data (still) exist?

  • Ethics: What are the ethical implications of data-mining and profiling? What is the link between prediction and prevention? Is there risk of self-fulfilling prophecy, when acting upon data-driven assumptions will turn prediction turn into reality? How can we preserve the individual right to self-determination? Do private actors have a duty or responsibility to act if they identify potential risk citizens? Should researchers and developers share the same ethical duties and responsibilities in the field of big data analytics? Are there any best-practices and guidelines in this field?

Submission guidelines

Extended abstracts should be between 1000 and 1500 words.

Papers’ length should be between 7000 and 10000 words and should be submitted in Microsoft Word or RTF format.

To provide a further stimulus to discussion, a scientific committee of renowned scholars will act as discussants for each of the presented papers.

Selected papers will be published in an edited volume.

Important Dates  

The following are the key dates for the conference:

  • Extended abstract by e-mail: May 2015, 15 (to either of the organisers)

  • Decision by organisers: May 2015
  • Completed papers: 1st of September 2015
  • Conference days: 30th and 31st of October 2015

Further Information

Up to date information on the workshop, as it becomes available, can be found at our website. Any important information will be mailed to you directly.

Our website can be found at:  http://blockchainworkshops.org/predict-prevent

If you have any questions about this conference, please contact us at:  

  • CharlotteCharlotte van Ooijen (CERSA/CNRS Université Paris II): charlotte.van-ooijen@cnrs.fr
  • PrimaveraPrimavera De Filippi (CERSA/CNRS Université Paris II): primavera.de-filippi@cnrs.fr

Scientific Committee

  • Christine Balagué (Conseil National du Numérique)
  • Pierre Bellanger (Skyrock)
  • Didier Bigo (Department of War Studies, King's College London; Sciences Po Paris)
  • Daniele Bourcier (CERSA-CNRS, Paris)
  • Max Dauchet (Commission d’Ethique, ALLISTENE)
  • Guillaume Farde (Spallian)
  • Olivier Gohin (Université Paris 2)
  • Paul De Hert (Law Science Technology & Society - LSTS, Vrije Universiteit Brussel; Tilburg Institute for Law, Technology, and Society - TILT, Tilburg University)
  • Deirdre Mulligan (Berkeley University)
  • Guillaume Piole (Supelec)
  • Olivier Renaudie (CERSA-CNRS, Paris)
  • Antoinette Rouvroy (Université de Namur)
  • Guillaume Sire (Institut Français de Presse)
  • C. William R. Webster (Centre for Research into Information, Surveillance and Privacy - CRISP; Stirling Management School, University of Stirling)

Lieux

  • Salle des Conseils, Université Paris II - Université Paris II 12, Place du Panthéon
    Paris, France (75005)

Dates

  • vendredi 15 mai 2015

Mots-clés

  • crime prevention, datamining, profiling, big data, quantified self, legitimacy, surveillance

Contacts

  • Primavera De Filippi
    courriel : primavera [dot] de-filippi [at] cnrs [dot] fr
  • Charlotte van Ooijen
    courriel : charlotte [dot] van-ooijen [at] cnrs [dot] fr

URLS de référence

Source de l'information

  • Charlotte van Ooijen
    courriel : charlotte [dot] van-ooijen [at] cnrs [dot] fr

Pour citer cette annonce

« Profile, Predict and Prevent », Appel à contribution, Calenda, Publié le mardi 21 avril 2015, http://calenda.org/325332