Facial Recognition in Policing: Hidden Dangers and What You Must Know

 Key Points

• Research suggests facial recognition technology (FRT) in law enforcement can misidentify individuals, particularly minorities, due to biased algorithms.

• It seems likely that India’s lack of strict privacy laws enables unchecked FRT use, risking mass surveillance.

• Evidence leans toward accountability gaps in India, with global efforts like the EU’s AI Act offering stronger oversight.

• Controversy surrounds FRT’s potential to violate privacy and amplify discrimination, with calls for ethical governance growing.

 Facial recognition technology (FRT) is changing how police solve crimes, but its ethical pitfalls—like bias, privacy violations, and lack of accountability—demand urgent attention. Let’s break down the risks, backed by real-world examples, and explore how countries like India, the U.S., and the EU are handling (or mishandling) this powerful tool.

1. Bias: When Technology Sees Skin Color, Not Facts

FRT isn’t neutral. Studies show it misidentifies people with darker skin tones or ethnic features more often, especially when trained on limited datasets.

India’s Delhi Riots (2020):

During the riots, Delhi Police used FRT to identify suspects. Reports revealed minorities, particularly Muslims, were wrongly flagged due to biased algorithms. The police considered an 80% accuracy rate “acceptable,” but experts argue this threshold is dangerously low [1],[2]. A study found Muslims in Delhi are disproportionately targeted, reflecting systemic biases [3].

Aadhaar’s Flawed System:

India’s biometric ID system, Aadhaar, integrates FRT, but poor-quality scans in rural areas often misidentify manual labourers (many from lower castes). These errors deepen inequality [4].

Global Parallels:

In the U.S., a 2018 MIT study found FRT error rates soared to 34.7% for darker-skinned women versus 0.8% for lighter-skinned men [5]. Robert Williams, a Black man, was wrongfully arrested in Michigan in 2020 after a faulty FRT match [6]. China uses FRT to target Uyghur minorities, enabling mass detentions [7].

2. Privacy: Who’s Watching You?

FRT collects sensitive facial data, often without consent. Weak laws in India enable unchecked surveillance, while other regions try (and sometimes fail) to balance security and rights.

India’s Surveillance Surge:

Delhi installed 1,000+ FRT-enabled cameras by 2022, yet no clear privacy rules exist to govern their use [8]. Despite a 2017 Supreme Court ruling declaring privacy a fundamental right, enforcement is lax [9]. The 2023 Digital Personal Data Protection Act (DPDP) exempts government agencies from consent requirements, raising fears of misuse [10].

EU vs. U.S.:

The EU’s GDPR requires explicit consent for biometric data and limits public-space FRT, though loopholes remain [11]. Meanwhile, the U.S. has no federal FRT laws—some cities ban police use, while others don’t, creating a patchwork of protections [12].

China’s Authoritarian Model:

FRT fuels China’s social credit system, prioritizing state control over privacy. Citizens have little recourse against surveillance [7].

3. Accountability: Who’s Responsible When FRT Fails?

When FRT makes mistakes, victims often have no way to seek justice. India’s opaque systems and lax corporate oversight worsen the problem.

India’s Opaque Deals:

Delhi Police buys FRT from private firms like Staqu without public audits. Proposed national systems, like the AFRS, centralize data without accountability [13],[14]. Companies like Facegram sell flawed tech to police, skipping bias checks [3].

Global Shifts:

In 2020, Microsoft, Amazon, and IBM paused FRT sales to police over ethical concerns [15]. The EU’s AI Act labels public-space FRT “high-risk,” mandating audits and transparency [16]. France fined a firm for restricting FRT access, promoting accountability [16].

4. Who Suffers Most? Marginalized Communities

FRT doesn’t impact everyone equally. In India, caste and religion play a role, while globally, refugees and minorities bear the brunt.

India’s Caste Divide:

Lower-caste groups (Dalits) and Muslims face heightened surveillance, as seen during the Delhi riots [3]. Rural areas with poor internet and cameras see more errors, hurting marginalized communities [4].

Global Inequity:

The EU uses FRT to profile migrants at borders [11]. In the U.S., predictive policing tools over-target Black neighbourhoods, worsening racial disparities [17].

5. How Can We Fix This?

Ethical FRT use is possible—with effort. Here’s what experts recommend:

Audit Algorithms:

Mandate public audits of FRT systems to catch bias, as proposed in the EU’s AI Act [16].

Diversify Data:

Train FRT on India-specific datasets (like IIT Madras’ “AI for Social Good”) to reduce errors [18].

Strengthen Laws:

Reform India’s DPDP Act to remove government exemptions and align with GDPR’s consent rules [10][11].

Global Collaboration:

Adopt UNESCO’s AI ethics guidelines to prioritize human rights and transparency [19].

The Bottom Line

FRT could enhance safety—but not without guardrails. India’s weak regulations and diverse population amplify risks, while the EU and U.S. grapple with their own challenges. By learning from global models (like audits and diverse datasets) and closing accountability gaps, we can ensure FRT serves justice—not injustice.

Stay informed. Demand transparency. Your face shouldn’t cost you your rights.

What’s Next for Facial Recognition and Public Safety?

As technology evolves, facial recognition tools are expected to become faster and more accurate. However, pressure is also rising for governments and companies to create stronger privacy laws, transparency rules, and ethical guidelines. Striking the right balance between security and civil rights will shape the future of how data science impacts policing across the world.

References

[1] Delhi Police's Use of Facial Recognition Technology
[2] Delhi Police's Facial Recognition System Concerns
[3] Facial Recognition for Policing in Delhi
[4] Internet Freedom Foundation on Aadhaar Errors
[5] MIT Study on FRT Bias
[6] ACLU on Wrongful Arrest
[7] China's Use of FRT
[8] Delhi's Surveillance Expansion
[9] Puttaswamy Judgment
[10] India's DPDP Act
[11] EU's GDPR Rules
[12] U.S. FRT Privacy Laws
[13] Delhi Police & Staqu
[14] National AFRS Risks
[15] Corporate Moratoriums
[16] EU's AI Act
[17] U.S. Predictive Policing
[18] IIT Madras' AI Initiative
[19] UNESCO's AI Ethics