iBorderCtrl video lie detector is branded 'hogwash'

Occurred: November 2018

The pilot of an EU-funded system that uses AI to detect lies at international airports was dismissed as 'Orwellian' pseudoscience.

In November 2018, CNN revealed that the EU was testing iBorderCtrl, a programme that featured a machine learning-based video 'lie detector module,' in Hungary, Greece and Latvia. The detector used an avatar of a border guard to ask people 13 questions about their personal backgrounds and travel plans, assessed travelers' micro-expressions, and notified human agents if it suspected them of being dishonest.

However, experts criticised the tool as pseudoscientific and likely to lead to unfair outcomes based on the inaccuracy and bias associated with many lie detection and facial recognition systems. iBorderCtrl was also described by privacy advocates as a 'dystopian' data collection vehicle, and an 'Orwellian nightmare'.

In July 2019, The Intercept found that the technology made several errors, incorrectly identifying four out of sixteen honest answers as false. 

Developer: European Dynamics
Country: European Union
Sector: Govt - immigration
Purpose: Detect traveller lies
Technology: Behavioural analysis; Emotion recognition; Facial recognition
Issue: Accuracy/reliability; Human/civil rights
Transparency: Governance