Background

While heralded for its automated decision-making possibilities, the advent of many Artificial Intelligence (AI) technologies poses a number of serious risks to fundamental rights, gender equality, and racial discrimination. The widespread use of facial recognition technologies has particularly alarming consequences for critical human rights issues, fuelling mass surveillance and racial profiling.

Facial recognition allows for the automatic identification of an individual by matching faces from digital images. Using biometric data, it can detect various facial features and compare them with other faces. It is already widely used in the private sector and, recently, has become more widespread within public administration, including law enforcement and border management agencies.

According to the Commission White Paper on AI and European data protection rules, the processing of personal data for identifying a natural person is prohibited, except under specific conditions. According to the General Data Protection Regulation (GDPR), data processing should mainly take place for reasons of substantial public interest. Under EU or national law, the processing of personal data, such as facial images, by public authorities should not infringe on fundamental rights.

Impact on Fundamental Rights

At the European level, there is no adequate research on the impact of facial recognition technologies on fundamental rights. However, in a new study, the United Nations Human Rights Council has found that ‘‘In the absence of robust rights protections which are institutionally embedded to oversee the collection, storage, and use of such evidence, relevant practices are likely to infringe international human rights law standards.’’

Facial recognition technologies will jeopardise a number of fundamental rights that are safeguarded under European law. For example, the respect for private life and personal data, and the preservation of human dignity will be central to the deployment of facial recognition technology in public spaces. The processing of facial images will violate a person’s right to privacy. However, the Commission has refused to take a stance on facial recognition and has repeatedly stated that individual member states should assess how and when they wish to permit the use of facial recognition technology and a ‘Europe-wide debate’ should be held on this issue.

However, the biggest problem with most algorithms used in facial recognition technologies is the misidentification of non-white or non-male faces. According to a test conducted in the US, most systems recorded false positive matches for Asian and African American groups. Inaccuracy and misidentification of facial recognition for non-white groups by law enforcement authorities can lead to wrongful arrests, charges, and detention. In Europe, a lack of regulation of such technologies will bring similar problems, exacerbating discrimination.

There is a worldwide call for clarity on this issue, with recent rulings that facial recognition breaches human rights law, and Black Lives Matter protests in the US forcing companies like Amazon and Google to adopt a voluntary moratorium until there is clear regulation. However, despite concerns from NGOs and civil liberties groups, the Commission has backed away from plans to ban or put a moratorium on facial recognition technologies by law enforcement agencies.

The View of the Left

To prevent an infringement of fundamental rights and guarantee privacy protection for all individuals, including migrants, women and people of colour, we believe that there should be a Europe-wide ban or moratorium on the use of facial recognition technology in public spaces. The Commission has made the development of AI and related technologies a top priority, focusing solely on innovation and attracting tech investment in Europe. However, as backed up by various studies on this issue, we call for a ban or a moratorium, as facial recognition is prone to inaccuracy, discrimination and identity fraud.

Moreover, as the Left group in the European Parliament, we demand a human-centred development of AI technology, in full compliance with the fundamental rights standards set out in law and the promotion of social inclusion, sustainability and fairness. A first step should be a ban on facial recognition as the only way to preserve our dignity and rights.