Meta’s AI smart glasses: systemic surveillance risks for marginalised communities under global tech colonialism

Indigenous epistemologies reject facial recognition as a form of epistemic violence, where data extraction severs relational knowledge systems (e.g., Māori data sovereignty, Andean communal consent models). Systems like Meta’s smart glasses replicate the colonial gaze, reducing human identity to extractable data points while ignoring Indigenous protocols of collective consent and reciprocity. The lack of Indigenous consultation in biometric tech design reflects a broader pattern of tech colonialism, where Western corporations and states impose surveillance infrastructures on Indigenous lands and bodies.

Facial recognition builds on 19th-century anthropometry and eugenics, where facial features were used to justify racial hierarchies (e.g., Bertillon’s mugshot systems, South Africa’s apartheid pass laws). The 20th century saw biometric IDs weaponised in Nazi Germany and colonial India, while today’s smart glasses echo 1980s 'smart card' experiments in authoritarian regimes like Singapore. Sex workers have faced surveillance-driven criminalisation since the 19th century, but modern AI systems scale this violence exponentially by automating identification and profiling.

In China, facial recognition is used to enforce social credit systems and persecute Uyghurs, while in Brazil, sex workers report being profiled by police using AI-enhanced surveillance tools. Indigenous communities in Canada and Australia have successfully blocked biometric data collection, arguing it violates treaty rights and cultural autonomy. African feminist tech collectives like *Afro-feminist AI* critique how facial recognition reinforces colonial gender norms by pathologising Black and queer bodies as 'deviant' or 'suspicious'.

Facial recognition systems exhibit racial and gender bias, with error rates up to 35% for darker-skinned women (NIST 2019), due to training datasets dominated by white male faces. Studies show AI surveillance increases police violence against marginalised groups (ACLU, 2020), while sex workers’ rights organisations document how facial recognition enables stalking and extortion. The 'solution' of 'anonymisation' is a myth: re-identification attacks can de-anonymise 99.98% of faces in supposedly anonymised datasets (Nature, 2021).

Artists like Zach Blas critique facial recognition as a tool of 'biometric capture,' using glitch art to disrupt algorithmic vision and reclaim bodily autonomy. Indigenous artists such as Hulleah Tsinhnahjinnie (Seminole/Muscogee) use photography to challenge colonial imaging practices, while queer and trans performers like Tourmaline use film to document surveillance’s impact on marginalised communities. Spiritual traditions like Buddhism warn against the 'illusion of the self' in data, where identity is reduced to a series of pixels rather than a dynamic, relational being.

By 2030, Meta’s smart glasses could enable real-time facial recognition in 1 billion devices, creating a global surveillance network where marginalised groups are preemptively profiled as 'high-risk.' Scenario modelling suggests this could trigger a backlash, with cities banning smart glasses (e.g., San Francisco’s 2019 ban on facial recognition) and sex workers developing counter-surveillance tools like 'face-blocking' makeup or decentralised identity systems. The tech could also accelerate authoritarianism, as seen in Russia’s use of smart glasses for protester identification.

Sex workers’ rights groups (e.g., SWOP, Global Network of Sex Work Projects) have long warned about AI surveillance, but their expertise is sidelined in tech policy debates. Trans and non-binary communities face heightened risks from facial recognition misgendering, yet their data is often used to 'train' biased algorithms without consent. Migrant workers in the Gulf States report facial recognition being used to track and deport them, while Black communities in the US describe it as 'digital stop-and-frisk.' These voices reveal surveillance as a tool of racial capitalism, not just a privacy issue.