Ravi, SP — AI-enabled predictive policing flagged as biased against an immigrant, low-income neighbourhood.
250 words across both parts, 20 marks. Identify stakeholders and the conflicting values, weigh the options, then commit to one course with safeguards. There is no single correct answer — there is a defensible one.
Common trap: Part (b) says 'place yourself in Ravi's role' — write in the first person and commit to one option. Note that Ravi commissioned the system himself: acknowledging that conflict and inviting external audit is what distinguishes a top-band answer.
The case
Ravi is a senior police officer with vast experience in riot control and cyber-policing. Since one year, he has been the Superintendent of Police (SP) of a district with a history of frequent rioting.
Last year, Ravi had sought installation of an AI enabled software for predictive policing. This system has been operational for approximately six months. This new system employs advanced algorithms for capturing the biometric data of persons in a crowd and swiftly relating it to a data library. This has enabled the police to identify the persons involved in various crimes.
The system has identified an immigrant and low-income neighbourhood as a centre for gang violence and drug trafficking. Aided by this AI analysis, the local police has focused its patrolling, preventive detentions and establishing checkposts. Consequently, public order and law enforcement has visibly improved.
Last week, some community leaders, civil rights lawyers and human rights activists visited Ravi's office. They submitted a memorandum that the new system is faulty as it is based on incorrect historical data caused by social biases and discriminatory policing. The memorandum also alleges that the increased surveillance has created a climate of tension amongst residents. This feeling is aggravated by the fact that the residents are not aware of the data noted against their names.
Model answer
20 marks · 250 wordsA measurable improvement in public order, achieved through a system that may be encoding past discriminatory policing as future prediction — with the affected residents unable to see or contest the data held on them.
| Residents of the flagged neighbourhood | Liberty, privacy, freedom from disproportionate stop-and-detain; no notice of data held against them |
| Ravi | Duty to maintain order; accountability for a system he commissioned; professional integrity |
| The wider district | Genuine security gains from reduced rioting and trafficking |
| Police personnel | Operating on algorithmic direction they cannot interrogate |
| Civil society and the courts | Constitutional oversight of surveillance and equal protection |
- Feedback loop bias — historical arrest data reflects where police patrolled, not where crime occurred; the model then directs patrols back to the same area, confirming its own prediction
- Proxy discrimination — even without using ethnicity as a variable, location and income can serve as proxies, producing indirect discrimination under Article 14 and 15
- Privacy — mass biometric capture of persons in a crowd, most of whom are not suspects, engages Puttaswamy's proportionality requirement and lacks a clear statutory basis
- Opacity and due process — residents cannot know, contest or correct the data recorded against them; there is no notice, hearing or appeal
- Presumption of innocence — preventive detention on predicted rather than committed conduct inverts it
- Automation bias — officers deferring to algorithmic output as objective, diluting individual professional judgement
- Accountability gap — if the system errs, responsibility diffuses between vendor, data and officer
- Consequentialist trap — improved order is real, but improvement achieved by over-policing a minority is not a neutral success
Course of action. Continue the system in a restricted mode with immediate safeguards, and simultaneously commission an independent technical and legal audit with community representation.
Why. Neither of the polar options is defensible. Dismissing the memorandum treats a serious equal-protection allegation as noise, while abandoning the system concedes a factual claim that has not been tested. The middle course is not a compromise but the ethically required step: establish whether the bias exists before deciding, and stop the clearly indefensible practices in the meantime. Predictive output should immediately cease to be sufficient ground for preventive detention, which must rest on individualised suspicion — this restores the presumption of innocence without waiting for the audit.
Safeguards. Suspend biometric capture of non-suspects pending legal review of its statutory basis and proportionality under Puttaswamy. Publish the categories of data held and create a notice, access and correction mechanism for residents. Require every algorithmic flag to be corroborated by independent human judgement, recorded in writing. Audit outcomes by neighbourhood for disparate impact. Establish a community liaison committee with the memorandum's signatories. Report the review honestly upward even if it embarrasses a system Ravi himself sought.
Verdict. The ethically optimal action is the audit-plus-safeguards route, because it is the only one that treats both the security gain and the bias allegation as facts to be established rather than positions to be defended. Ravi's personal authorship of the system is precisely why he must invite external scrutiny rather than conduct the review himself.
More from GS Paper 4 (Ethics) 2026
Model answer by the EdiShorts editorial team. It is one well-argued treatment of the question, written to the official word limit — not an official UPSC answer key.