Scrutinizing Surveillance Systems: Modern Technologies of Security, Gender Biases and Biopower

The use of modern technology skyrocketed in the late twentieth century, and we never looked back. As the world has become more reliant on technology for day-to-day activities and ensuring safety and security, some innovations need to be scrutinized further in the context of international security and what it truly means to be secure. One such cog in the machine of the critical infrastructure that drives security efforts within the sovereign bounds of the state is artificial intelligence-driven (AI) and independent surveillance systems. Under the pretense of heightening security for the safety of the individual, states have engaged in the uncontrolled expansion of surveillance networks.

This blog post aims to uncover how securitization, feminist critiques and biopower work to form new conceptions of security in surveillance technology powered by AI.

The Copenhagen School would fail to account for increased undue surveillance as an ‘existential threat’ to the referent object, allowing surveillance schemes to seize control of individual liberties (Hansen, 2000). Johan Galtung and Jan Øberg’s conception of security issues somewhat provides a solution by including threats to an individual's survival, development, freedom, and identity (Waever, 1995). In an attempt to broaden the scope of security, Ole Waever’s conception of security, which recognises ‘state power claiming the legitimate use of extraordinary means’ as a valid threat to security, encompasses AI surveillance mechanisms perpetuated by the state (Weaver, 1995). 

Artificial intelligence (AI) based surveillance systems are riddled with pathologies, a prominent one being the presence of gender bias. Experiments and observation have revealed that a gender bias plays out in these systems, with significantly lower accuracy for female faces than men (Albiero et al., 2022). But, for now, AI is not autonomous; it is predictable.

The most prominent explanation for this pathology of gender bias is intricately linked to what Hansen points out as the absence of gender or, in this case, an imbalanced dataset (Hansen, 2000).

Artificial intelligence-based surveillance systems using facial recognition technology are trained on extensive data sets that attempt to simulate every permutation of a face the system could encounter in the future. When the dataset does not contain enough instances of female faces, the system becomes worse at accurately recognising these faces. The intersectionality of identities and overlapping multiple underrepresented characteristics in training data exacerbates this problem. Training data is not sensitised to gender identities, leading to faulty classification and threats to the individuals' personal security (Albiero et al., 2022). The state-led nature of these profiling mechanisms makes it impossible to voice out security concerns for fear of aggravating the threat, coinciding with Hansen's framework of 'security as silence' (Hansen, 2000).

The increase in surveillance methods has inevitably led to a higher concentration of power in the hands of the sovereign state. In this context, Michel Foucault’s concepts of governmentality and biopolitics go hand in hand to uncover the sinister reality behind state-sponsored surveillance mechanisms (Sokhi-Bulley, 2021). Extensive collection of biometric data like fingerprints, iris and retina scans and voice recognition, among others, has become a mainstay of every sovereign state in the context of high levels of digitisation of government administration processes.

Take the example of India, which has proliferated surveillance systems under the guise of developmental projects like ‘Digital India’, ‘Viksit Bharat’ and ‘Smart Cities’. These projects, along with the central Aadhar repository, give the state unparalleled ‘biopower’ or control over the lives of Indian citizens (Mbembe, 2003). For the citizens, heightened ‘security’ measures have only resulted in apathy.

The illusion of securitisation by surveillance through the ‘strategic use of technology to maximise control over populations’ has led to the zombification of the average citizen, who is unaware of where their data is being recorded, sent and compiled (ibid). 

For the citizens of a state that increasingly begins to look like a panopticon, with cameras pointing at them from every direction, insecurity sets in. Mbembe's analogy of slave life, on the brink of death, given bare subsistence to ensure productivity appears much like the modern techno-capitalist state system (ibid).

---

REFERENCES 

Hansen, L. (2000, June). The Little Mermaid’s Silent Security Dilemma and the Absence of Gender in the Copenhagen School. Millennium: Journal of International Studies, 29(2), 285–306. https://doi.org/10.1177/03058298000290020501

Albiero, V., Zhang, K., King, M. C., & Bowyer, K. W. (2022). Gendered Differences in Face Recognition Accuracy Explained by Hairstyles, Makeup, and Facial Morphology. IEEE Transactions on Information Forensics and Security, 17, 127–137. https://doi.org/10.1109/tifs.2021.3135750

Waever (Ed.). (1995, October 13). Securitization and Desecuritization. In On Security. https://doi.org/10.1604/9780231102704

Sokhi-Bulley, B. (2021, July 8). Governmentality: Notes on the Thought of Michel Foucault. Critical Legal Thinking. https://criticallegalthinking.com/2014/12/02/governmentality-notes-thought-michel-foucault/

Mbembe, A. (2003). Necropolitics. Translated by Libby Meintjes. Public Culture, 15(1).