Go back to the Europarl portal

Choisissez la langue de votre document :

  • bg - български
  • es - español
  • cs - čeština
  • da - dansk
  • de - Deutsch
  • et - eesti keel
  • el - ελληνικά
  • en - English (Selected)
  • fr - français
  • ga - Gaeilge
  • hr - hrvatski
  • it - italiano
  • lv - latviešu valoda
  • lt - lietuvių kalba
  • hu - magyar
  • mt - Malti
  • nl - Nederlands
  • pl - polski
  • pt - português
  • ro - română
  • sk - slovenčina
  • sl - slovenščina
  • fi - suomi
  • sv - svenska
Parliamentary questions
PDF 39kWORD 10k
10 January 2020
E-000152/2020
Question for written answer E-000152/2020
to the Commission
Rule 138
Patrick Breyer (Verts/ALE)
 Answer in writing 
 Subject: iBorderCtrl: False incrimination by and discriminatory effects of video lie detector technology

The iBorderCtrl project has piloted a prototype of the ‘Intelligent Border Control’ system. An Italian journalist reported that of the 16 questions she truthfully answered when participating in the trial, the ‘deception detection’ component falsely assumed that she was lying four times. (1) .

1. On average, what proportion of participants were flagged as likely ‘liars’ by the ‘deception detection’ component, and what was the proportion of false positives among all answers identified by the ‘deception detection’ component as probable ‘lies’?

2. Are certain groups of people (e.g. people of colour, women, the elderly, children, people with disabilities) more likely than others to be falsely reported as liars by the ‘deception detection’ technology?

3. Why have the full results of the pilot tests not yet been published and when will they be published?

(1)https://theintercept.com/2019/07/26/europe-border-control-ai-lie-detector/
Last updated: 25 September 2020Legal notice - Privacy policy