Facial recognition and algorithmic bias in large Brazilian cities

• Rodrigo Brandão USP http://orcid.org/0000-0003-0127-4989
• João Lucas Oliveira USP

DOI: https://doi.org/10.5753/wics.2021.15970

Abstract

Are Brazilian governments aware of the social risks of facial recognition applications when they use them? To gather elements that allow us to answer this question, we investigated the digital official diaries of 13 of the 17 Brazilian cities with more than one million inhabitants. Based on the collected material, we preliminary analysed the use of facial recognition in the sector of public transport. We found that some local governments seem to be better prepared than others to deal with the risks in question: in the face of alleged cases of fraud, they allow the user to follow his/her journey, and their legal documents bring minimum guidelines on the role of human work in the process of fraud revision.

Keywords: artificial intelligence, facial recognition, bias, public sector

References

Berryhill, J. et al. (2019). “Hello, World: Artificial intelligence and its use in the public sector”. OECD Working Papers on Public Governance No. 36.

Brandão, R. et al. (2021). “Reconhecimento facial e discriminação algorítmica nos municípios brasileiros”, 7 de maio de 2021. Disponível em: https://www.migalhas.com.br/depeso/345092/reconhecimento-facial-e-discriminacao-algoritmica-nos-municipios

Buolamwini, J., Gebru, T. (2018). “Gender Shades: Intersectional Accuracy Disparities in Commercial Gender Classification”. Proceedings of Machine Learning Research 81:1–15. Conference on Fairness, Accountability, and Transparency.

Chaudhuri, B. (2019). “Paradoxes of Intermediation in Aadhaar: Human Making of a Digital Infrastructure,” Journal of South Asian Studies 42 (2019):572–587.

Coelho, J., Burg, T. (2020). “Uso de inteligência artificial pelo poder público”. Transparência Brasil.

Eubanks, V. (2018). Automating Inequality. St.Martin's Publishing Group. Edição Kindle.

Kak, A. (2020). “The State of Play and Open Questions for the Future”. In: Regulating Biometrics – Global Approaches and Urgent Questions, Editado por Amba Kak, IA Now, Estados Unidos da América.

MIT Technology Review. (2019). “A US government study confirms most face recognition systems are racist”, 20 de dezembro de 2019. Disponível em: https://www.technologyreview.com/2019/12/20/79/ai-face-recognition-racist-us-government-nist-study/ [consultado em 20-02-2021].

Shearer, E., Stirling, R., Pasquarelli, W. (2020). “Government AI Readiness Index 2020”. IDRC & Oxford Insights.