The Use of Facial Recognition Technology in EU Law Enforcement: Fundamental Rights Implications

Thumbnail Image
Dushi, Desara
Journal Title
Journal ISSN
Volume Title
Global Campus of Human Rights
Facial recognition technology is a type of biometric application used to identify people’s faces based on datasets and then makes assessments about those people based on algorithmic predictions. This technology can be used for three types of analytics: verification (matching the ID photo in airports), identification (matching a photo in a database) and classification (gender, age, etc). This technology is widely used by private companies for advertisement and marketing, by analysing facial expressions of clients to predict their preferences; for identifying ideal job candidates; or for automatic tagging of people in photos (Facebook for example). But, facial recognition is not used only by the private sector. Its evolution has attracted the public sector too, especially law enforcement and border management. This has generated many debates on the impact on human rights. Artificial Intelligence (AI) systems are typically trained on data generated by people. Therefore, it is possible that any AI system would reflect the social biases of the people who developed their datasets. On the other hand, it raises concerns of breach of privacy when used in public spaces (ie mass surveillance), discrimination (the algorithm has proven problematic for people of colour), false labelling based on facial expressions (ie in interviews or for criminal profiling), unwanted tagging and when used to send advertisements based on shops people have visited. It also causes intimidation to people and a feeling of intrusiveness. Public safety and expression of consent by people are classic justifications behind the use of such identification technology. But questions remain: Is it necessary? is it the best/right remedy? is it proportional? is it effective? and, ultimately, is the expressed consent informed consent?
European Union, human rights, boundaries, surveillance, technological innovation, discrimination, privacy, consent, recognition