Latin American city governments rolling out facial recognition in public spaces for law enforcement purposes are failing to properly assess potential human rights impacts and offer weak protections to citizens exposed to the technology.
When the City of Buenos Aires first deployed live facial recognition in 2019, its deputy mayor asserted that ‘police forces [could] now do their job more efficiently’. With the technology being initially rolled out across the city’s rail networks, Argentina’s minister of transport further reinforced this point, highlighting that while ‘82% of passengers report[ed] feeling safe on trains, this technological innovation [would] make them feel even safer’.
The deployment of facial recognition technology in Latin America has been underpinned by claims about its ability to effectively identify potential criminals and enhance public safety. Growing security concerns in large urban centres, astute outreach by – in many instances foreign – surveillance technology companies and a marked tendency towards techno-solutionism on the part of both politicians and technocrats have led many local governments to buy into the potential capabilities of facial recognition. Other governments – especially those of an authoritarian nature – have been enticed by the opportunity for greater social control.
While the adoption of facial recognition has sparked widespread debate across the US and Europe, the technology has had a swift take-up across Latin America. Argentina, Bolivia, Brazil, Colombia, Ecuador, Mexico, Paraguay and Peru have ongoing deployments or have run trials. Chile, El Salvador, Guatemala and Uruguay have plans for its implementation, with Guatemala and Uruguay already running tenders for the purchase of the technology. Deployments are suspected to have begun in Honduras and Nicaragua.
Worldwide, multiple countries are beginning to grapple with how to regulate facial recognition. The European Union has proposed to treat biometric identification systems as a high-risk application of artificial intelligence (AI) and ban their use in public spaces, except for specific criminal investigations. In 2020, a British court found that the South Wales Police – the body in charge of running facial recognition trials in the UK – did not have sufficiently clear guidelines governing the deployment of facial recognition equipment in public places, and could have done more to assess the potentially discriminatory use of the technology. In the US, widespread city-level bans throughout 2019 and 2020 sparked a debate about whether the technology is compatible with fundamental rights. To date, five states – California, Maine, Massachusetts, New York and Vermont – have either placed moratoriums on the use of facial recognition or have strictly regulated its use for policing purposes, allowing its deployment in public spaces only under specific circumstances.
While Europe, the UK and the US are exploring regulatory frameworks to balance the tensions between the use of facial recognition in public spaces and the potential threats it poses to privacy and other fundamental rights, Latin America may well be described as being stuck in a worst-case scenario, in that the technology is frequently being deployed or piloted on the continent despite a lack of robust regulatory frameworks to ensure proper protections, oversight and transparency. Bans on the use of the technology do not appear to be options available for debate.
This research paper will focus on the use of facial recognition technologies in public spaces and for law enforcement purposes, with reference to the city-level deployments in the Argentinian capital, Buenos Aires, between 2019 and 2022, and the technology piloting in São Paulo, Brazil’s largest city, in 2020. Two potential routes emerge as alternatives to break away from the current worst-case scenario observed in the region. One route is the adoption of moratoriums on the use of facial recognition until proper safeguards are put in place and some uses of the technology possibly circumscribed. An alternative route, as proposed by human rights groups, is the enactment of comprehensive bans that prohibit the use of facial recognition for law enforcement purposes in public spaces.
Chapter 2 of the paper will set the scene, introducing the controversies surrounding the deployment of facial recognition, and the potential human rights impacts of the technology. Chapter 3 will look at trends in facial recognition deployments in Argentina and Brazil, focusing on the cases of Buenos Aires and São Paulo. Through the study of these deployments, the paper will document the underlying dynamics that are at play, as well as common trends in the adoption of this specific surveillance technology in the region. Chapter 4 will pose the question of why deployments in Latin America continue to take place despite negative human rights impacts. It suggests that politics plays a central role in motivating facial recognition rollouts. Drawing on policy developments in other jurisdictions, Chapter 5 will explore possible approaches to regulating facial recognition in Latin America, and Chapter 6 will present a concluding discussion on whether to ban or to regulate facial recognition in Latin America, offering a series of recommendations for regional policymakers.
