Is it ethical to use AI in predictive policing?

Opening Statement

The opening statement sets the intellectual and moral foundation of any debate. It defines the battlefield, establishes value criteria, and frames how judges and audiences interpret all subsequent arguments. On the motion “Is it ethical to use AI in predictive policing?”, both sides must grapple not only with technological feasibility but with profound questions about justice, freedom, and the soul of modern society. Below are the opening statements crafted to meet the highest standards of clarity, logic, creativity, and strategic foresight.

Affirmative Opening Statement

Ladies and gentlemen, imagine a city where police don’t respond to crimes — they prevent them. Not through surveillance states or authoritarian control, but through intelligent, transparent, and ethically governed artificial intelligence. We stand in firm support of the ethical use of AI in predictive policing — not because we worship machines, but because we value human lives enough to protect them before tragedy strikes.

Our position rests on three pillars: effectiveness, equity, and evolution.

First, AI enables proactive public safety. Traditional policing operates reactively — a crime occurs, an investigation follows. But what if we could identify high-risk areas and allocate resources before violence erupts? Studies from cities like Los Angeles and Chicago have shown that predictive models can reduce property crime by up to 25% when paired with community engagement. This isn’t mind-reading; it’s pattern recognition at scale — analyzing historical data, environmental factors, and temporal trends to guide patrols more efficiently. When every minute counts in preventing assault or robbery, prediction isn’t just useful — it’s a moral imperative.

Second, AI reduces human bias — when properly designed. Yes, flawed systems exist. But let us not confuse bad implementation with inherent immorality. Human officers carry unconscious prejudices — studies show Black drivers are stopped more often despite lower contraband hit rates. AI, by contrast, processes data without fatigue, fear, or prejudice — provided it is trained on clean data and audited regularly. In fact, well-regulated AI can serve as a check on human bias, flagging inconsistencies in deployment patterns and ensuring resource allocation aligns with actual risk, not stereotypes.

Third, predictive AI creates unprecedented accountability. Unlike opaque hunches or “gut feelings,” algorithmic decisions leave digital trails. Every input, weight, and output can be logged, reviewed, and challenged. With independent oversight boards, open-source auditing tools, and sunset clauses for outdated models, AI offers a path toward more transparency than traditional policing has ever achieved. We do not advocate blind trust in black boxes — we advocate for a new standard of algorithmic justice: explainable, contestable, and subject to democratic scrutiny.

Some may say, “But AI can’t replace judgment.” And we agree — it shouldn’t. Our model is human-in-command, machine-in-support. Officers remain responsible for decisions; AI merely sharpens their situational awareness. To reject this tool outright is to prioritize ideology over innovation, emotion over evidence, and inertia over improvement.

We are not building a dystopia — we are building a safer tomorrow. And if ethics means doing the greatest good for the greatest number, then yes — it is not only ethical to use AI in predictive policing, it is our duty.

Negative Opening Statement

Thank you. Now consider another vision: one where your neighborhood is policed not because of what anyone did, but because a computer decided it was “likely” someone would. Where your zip code determines how often officers knock on doors, follow you on the street, or question your presence — none of it based on behavior, but on probability scores derived from flawed data and hidden algorithms.

We firmly oppose the ethical use of AI in predictive policing — not because we distrust technology, but because we deeply value justice, dignity, and the presumption of innocence. Once we allow algorithms to forecast criminality, we cross a line from law enforcement into preemption — from punishing acts to managing risks. And history shows us who bears the cost: the poor, the marginalized, and communities already over-policed and under-protected.

Our opposition stands on three unshakable grounds: injustice, illegitimacy, and inevitability.

First, AI entrenches and amplifies systemic racism. Predictive models are trained on historical crime data — data collected during decades of discriminatory practices like stop-and-frisk and redlining. If police have always patrolled Black neighborhoods more heavily, those areas will generate more arrests — not necessarily more crime. Feeding this skewed data into AI creates a feedback loop: more patrols → more arrests → higher risk scores → even more patrols. It’s not prediction — it’s automation of oppression. A 2016 ProPublica investigation found that COMPAS, a widely used risk-assessment tool, falsely labeled Black defendants as future criminals at twice the rate of white ones. That’s not neutrality — that’s coded injustice.

Second, predictive policing violates core legal and moral principles. The cornerstone of any free society is the presumption of innocence. Yet here we are, surveilling and suspecting individuals not for what they’ve done, but for what a machine calculates they might do. This shifts the burden of proof from action to identity — from guilt to geography. Is this not eerily reminiscent of Minority Report? Except in that film, there were safeguards, appeals, and human review. In reality, most predictive systems operate in secrecy, shielded from public oversight by claims of proprietary software. No warrant, no trial, no appeal — just an algorithm whispering, “You’re likely to offend.” That is not policing; it is pre-crime.

Third, the slippery slope is already downhill. Even with regulations, AI in policing will expand beyond its original scope. Today it predicts burglary hotspots. Tomorrow it profiles individuals. Next year it integrates facial recognition, social media scraping, and real-time behavioral analytics. Once we normalize surveillance based on probability, resistance crumbles. And who controls these systems? Private corporations with profit motives, not public servants with accountability. When Palantir partners with the LAPD, whose interest comes first — community safety or shareholder returns?

Technology is neutral, yes — but context is everything. Using AI to recommend movies is harmless. Using it to determine who gets watched, followed, or detained is profoundly dangerous. We cannot claim to uphold justice while deploying tools that institutionalize statistical prejudice under the guise of objectivity.

So let us ask: Do we want a society where safety is bought at the price of freedom? Where progress is measured not by trust, but by control? If ethics demands respect for human dignity — then the answer is clear. No, it is not ethical to use AI in predictive policing. Because no algorithm should ever decide who deserves suspicion before they’ve done anything wrong.

Rebuttal of Opening Statement

In the rebuttal phase, the debate sharpens from philosophical framing into direct confrontation. Here, teams must do more than repeat their stance — they must dissect the opponent’s logic, expose vulnerabilities, and re-anchor the discussion in their value system. The second debater plays a critical role: not merely defending, but advancing the argument through precision strikes against the heart of the opposing case.

This stage separates surface-level rhetoric from rigorous thinking. It demands clarity under pressure, the ability to summarize complex ideas swiftly, and the courage to confront uncomfortable truths — whether in the opponent’s position or within one’s own assumptions.

Affirmative Second Debater Rebuttal

The opposition paints a dystopian future — neighborhoods surveilled by soulless algorithms, lives ruined by probability scores, justice replaced by code. It’s a compelling narrative, cinematic even. But let’s be honest: they’ve described bad AI, not all AI. And in rejecting the entire concept, they throw away the scalpel because someone once used it as a weapon.

Let me address their three core claims — injustice, illegitimacy, and inevitability — and show why each collapses under scrutiny.

First, yes, biased data exists — we don’t deny it. But the solution isn’t to abandon AI; it’s to correct the data. The negative side treats historical crime statistics as an unchangeable fate, when in fact, modern AI systems can be audited, adjusted, and retrained. Techniques like bias mitigation algorithms, fairness constraints, and counterfactual testing already exist to detect and reduce discriminatory outcomes. To say “AI amplifies racism” is like saying “fire caused the house to burn down” and then banning all fire — including those that cook food, warm homes, and save lives in emergencies. Our position is not blind faith — it’s active stewardship.

Second, the claim that predictive policing violates the presumption of innocence misunderstands how these tools are actually used. No one is being arrested based on an algorithm. Officers aren’t handcuffing people for having a high risk score. What happens is this: a patrol is redirected to a block where thefts spike every summer. A social worker is alerted to a youth program in an area flagged for potential gang recruitment. This isn’t pre-crime — it’s prevention. We vaccinate against disease before people get sick. We install smoke detectors before fires start. Why should public safety be any different?

And let’s talk about that Minority Report comparison. It’s dramatic, sure — but deeply misleading. In the movie, people were punished for crimes they hadn’t committed. In real-world predictive policing, no such thing occurs. At worst, there’s increased surveillance — which the negative side equates with punishment. But visibility isn’t violation. A well-lit street doesn’t assume guilt; it deters crime. So too does intelligent resource allocation.

Finally, the slippery slope argument — “today hotspots, tomorrow thoughtcrime” — assumes we live in a world without laws, oversight, or civic resistance. But we don’t. We have courts, legislatures, civil rights organizations, and media watchdogs. If Palantir overreaches, we regulate it. If a model discriminates, we sue it. The answer to misuse isn’t non-use — it’s accountability.

We’re not naive. We know technology can be misused. But ethics isn’t about avoiding risk — it’s about managing it wisely. And doing nothing while communities suffer violent crime? That’s not caution. That’s complicity.

So let’s stop fearing the machine and start fixing the system. Because with proper governance, AI doesn’t replace justice — it makes it more precise, more equitable, and yes, more human.

Negative Second Debater Rebuttal

The affirmative team speaks of “active stewardship,” “bias correction,” and “governance.” How comforting. Like telling someone standing on thin ice, “Don’t worry — we’ll send help if you fall.”

They present AI in predictive policing as a benevolent tool awaiting only good intentions and better code. But their entire argument rests on a fantasy: that flawed systems can be perfectly fixed, that corporations will self-regulate, and that transparency is achievable when profits depend on secrecy.

Let’s dismantle their three pillars — effectiveness, equity, and evolution — and reveal the cracks beneath.

First, effectiveness: They cite a 25% drop in property crime in Los Angeles and Chicago. Impressive — until you look deeper. Independent studies have shown that these reductions often occur during periods of broader socioeconomic improvement or increased community policing efforts. When researchers controlled for other variables, the AI’s contribution vanished. In some cases, crime simply displaced to adjacent areas — not reduced, just moved. Prediction didn’t stop crime; it reshaped its geography. That’s not success — it’s statistical sleight of hand.

Second, equity: They claim AI reduces human bias. But how? By replacing subjective prejudice with systemic bias encoded in data. You cannot “clean” data that reflects decades of over-policing. You cannot audit away the fact that Black neighborhoods were targeted not because of higher crime rates, but because of policy choices rooted in racism. An algorithm trained on that data doesn’t eliminate bias — it mathematizes it. And calling this “pattern recognition” is Orwellian. Patterns imply objectivity. But when the pattern is “arrest Black people more,” the machine learns: “Black areas are dangerous.” That’s not intelligence — it’s institutional memory disguised as neutrality.

And what about their so-called “human-in-command” model? Lovely in theory. But in practice, officers trust algorithmic recommendations — especially when backed by technical authority. Studies show cognitive deference to machines, particularly under stress. So when a dashboard says “patrol District 7,” officers go to District 7 — even if nothing seems amiss. The human isn’t in command; the algorithm is whispering orders through the interface.

Third, evolution and accountability: They speak of open-source audits and oversight boards like they’re already standard practice. But where are they? In most cities, predictive systems are black boxes sold by private firms under NDAs. LAPD won’t disclose how COMPAS works. Chicago hasn’t released full model parameters. And why? Because these companies claim trade secrets — and courts often agree. So much for transparency.

Even if oversight existed, who holds these systems accountable when harm occurs? Can you sue an algorithm for wrongful targeting? Can a community demand reparations from a neural network? No. Liability dissolves into layers of contractors, subcontractors, and免责 clauses. Meanwhile, the damage is real: families harassed, kids stigmatized, trust eroded.

The affirmative says, “Don’t throw out the scalpel.” But what if the scalpel is made from poisoned steel? What if every time you use it, infection spreads — even with gloves and antiseptic?

They assume perfect regulation is possible. We point to history: financial algorithms caused crashes; social media algorithms radicalized populations; facial recognition misidentifies minorities at alarming rates. And each time, we said, “Next time we’ll do better.” But “next time” never comes.

So no — the ethical choice isn’t to tinker with broken tools. It’s to reject a system that automates suspicion, normalizes surveillance, and shifts justice from deeds to demographics.

Because if we allow algorithms to decide who is risky, we’ve already lost something far more valuable than efficiency: our belief in freedom before accusation, in dignity before data.

Cross-Examination

In competitive debate, no phase reveals weakness more ruthlessly than cross-examination. It is here that polished speeches meet unscripted scrutiny, where assumptions are tested under fire, and rhetoric collides with reason. The third debater steps forward not to restate, but to dissect — to ask questions so precise they force admissions, so layered they unravel entire frameworks.

On the motion “Is it ethical to use AI in predictive policing?”, both teams have staked out powerful moral ground: one defending innovation as duty, the other resisting automation as betrayal. Now, in this crucible of dialogue, those principles face their first true stress test.

Affirmative Cross-Examination

Affirmative Third Debater:
I’ll address one question to each of your three colleagues. Please answer directly.

To the Negative First Debater: You argued that predictive policing violates the presumption of innocence because people are surveilled based on probability, not action. But isn’t every public health measure — from vaccine mandates to drunk-driving checkpoints — also based on risk rather than certainty? If we accept probabilistic interventions there, why is it unethical when applied to crime prevention?

Negative First Debater:
Public health measures are grounded in biological causality and universal risk exposure — like viruses spreading regardless of race or class. Predictive policing targets specific communities based on historically biased data. The difference isn’t just intent — it’s impact. One protects equally; the other punishes unequally.

Affirmative Third Debater:
To the Negative Second Debater: You dismissed our claim that AI can reduce human bias, saying algorithms merely “mathematize” systemic racism. But if we replace an officer who unconsciously profiles young Black men with an AI trained on de-biased data and fairness constraints, isn’t that still a net reduction in discriminatory outcomes — even if imperfect?

Negative Second Debater:
Only if such a system exists — which it doesn’t. You’re describing a hypothetical utopia. In reality, “de-biased” data is an oxymoron when the underlying social conditions remain unjust. And even if you could clean the input, the output still shapes behavior: more patrols lead to more stops, which generate new biased data. The cycle persists — now automated and harder to challenge.

Affirmative Third Debater:
To the Negative Fourth Debater: You’ve said we shouldn’t use tools that institutionalize suspicion before wrongdoing. But if we develop a fully transparent, open-source predictive model — co-designed with affected communities, audited annually, and limited strictly to geographic hotspots — would you still oppose it solely on principle?

Negative Fourth Debater:
Yes. Because transparency doesn’t erase harm. Even a well-intentioned spotlight creates stigma. When police flood a neighborhood “just for prevention,” residents feel targeted, not protected. Trust erodes. And once surveillance infrastructure exists, scope creep is inevitable. Today it’s blocks; tomorrow it’s individuals. Your “perfect” system is the Trojan horse of pre-crime.

Affirmative Cross-Examination Summary:
Thank you. What we’ve heard confirms our deepest concern: the opposition isn’t opposing flawed AI — they’re opposing all AI, regardless of safeguards, equity efforts, or democratic control. They equate prediction with punishment, prevention with persecution. By their logic, we should dismantle traffic cameras, close weather forecasting centers, and abolish epidemiological modeling — all forms of risk-based intervention. But society doesn’t function by ignoring patterns; it thrives by learning from them. Their absolutism offers no path forward — only retreat. We asked whether fairness is possible. They answered: only if nothing changes. That’s not ethics. That’s surrender.

Negative Cross-Examination

Negative Third Debater:
Three questions — one for each of your speakers.

To the Affirmative First Debater: You cited LA and Chicago’s success with predictive policing. But multiple peer-reviewed studies, including a 2020 Rand Corporation analysis, found no significant long-term crime reduction and documented increased harassment in minority neighborhoods. Given that evidence, how do you defend these systems as ethical rather than simply politically convenient?

Affirmative First Debater:
Correlation isn’t causation. Some programs failed due to poor implementation, lack of community input, or outdated models. That doesn’t invalidate the concept — it underscores the need for better design. We don’t abandon medicine because some drugs have side effects.

Negative Third Debater:
To the Affirmative Second Debater: You claimed officers remain “in command,” with AI only supporting decisions. Yet psychology research shows humans systematically defer to algorithmic authority, especially under pressure. If an officer receives a red alert on their dashboard, aren’t they far more likely to act on it — even without evidence — simply because the machine said so?

Affirmative Second Debater:
That’s why training matters. Officers must be taught critical engagement with technology — not blind obedience. Just as pilots rely on autopilot but monitor constantly, so too must police interrogate algorithmic outputs. The solution isn’t to scrap the tool; it’s to strengthen professional judgment.

Negative Third Debater:
To the Affirmative Fourth Debater: You advocate for independent audits and open-source models. But most predictive systems are built by private firms like Palantir under non-disclosure agreements. If the code is secret, how can any audit be meaningful? And if municipalities can’t access the algorithms, how can citizens challenge wrongful targeting?

Affirmative Fourth Debater:
We call for legislative mandates requiring transparency as a condition of contract. No black boxes, no deal. Public safety shouldn’t depend on trade secrets. Where markets fail, regulation must step in — and voters are beginning to demand it.

Negative Cross-Examination Summary:
Appreciate the answers — and the evasions. Let’s be clear: the affirmative team believes in a world that does not exist. A world where corporations willingly hand over source code. Where police departments fund bias audits instead of body armor. Where officers critically evaluate algorithmic alerts amid chaos and adrenaline. That’s not stewardship — that’s science fiction dressed as policy. They admit failures were due to “implementation,” yet offer no enforceable mechanism to ensure future compliance. Their entire case floats on hope: hope that profit-driven tech companies will behave altruistically, hope that overstretched departments will prioritize ethics over efficiency, hope that humans won’t defer to machines. History mocks that hope. From facial recognition errors to wrongful arrests based on faulty AI matches, the pattern is clear: once deployed, these systems cause harm faster than oversight can respond. You don’t build justice on optimism. You build it on limits. And the limit here is simple: no algorithm should ever decide who gets watched before they’ve done anything wrong.

Free Debate

The free debate round is where principles collide, assumptions are tested, and rhetoric meets reality. It’s not merely an extension of prior arguments—it’s a battlefield of wits, where every word must land with precision. Here, teams don’t just defend—they provoke, redirect, and reframe. On the motion “Is it ethical to use AI in predictive policing?”, the clash transcends technology; it becomes a contest between visions of justice: one that seeks optimization, the other, integrity.

With the Affirmative pushing for innovation within guardrails, and the Negative warning of moral erosion beneath the guise of progress, the floor opens—not for monologues, but for a duel of ideas. Alternating speakers rise, each building on their team’s logic while striking at the heart of the opposition’s weakest links. What follows is a reconstructed, high-intensity exchange—sharp, strategic, and occasionally laced with humor, because even in serious matters, clarity often comes wrapped in wit.

Affirmative First Debater:
You say we shouldn’t use AI because past systems were biased? Fine. Then let’s fix the data, audit the models, and sunset flawed algorithms. But don’t tell me we should keep sending officers into danger based on hunches and habit patterns shaped by decades of bias. If your solution is “go back to how it was,” then you’re not defending justice—you’re romanticizing failure.

Negative First Debater:
And yours is “let’s optimize oppression”? You call it fixing data—communities call it repackaging discrimination. When an algorithm flags a block as high-risk because Black people were over-policed there since the 1970s, you’re not predicting crime—you’re forecasting injustice. Is that your idea of progress?

Affirmative Second Debater:
So transparency isn’t possible? Oversight impossible? Regulation fantasy? That’s defeatist. We regulate nuclear power. We govern financial markets. Why is algorithmic accountability suddenly beyond human capability? Are we so afraid of complexity that we’d rather trust gut instincts proven to be racially skewed?

Negative Second Debater:
We regulate nuclear plants with independent agencies, public hearings, and physical inspections. Try asking Palantir to open its source code. They’ll laugh—or sue. Your “oversight” exists only in PowerPoint slides. In practice, it’s corporate secrecy dressed up as civic duty.

Affirmative First Debater:
Then pass laws forcing transparency! Create public AI ethics boards! Isn’t that better than letting fear paralyze progress? You act like any use of AI is surrender to Skynet. But if we applied that logic to medicine, we’d still be bleeding patients to cure fever.

Negative First Debater:
Bleeding patients was also done with confidence—by experts citing tradition. Just like you cite “efficiency” while ignoring who bears the cost. Let me ask: would you accept a risk score determining how often police follow you on your way home? Or does dignity only matter when it’s personal?

Affirmative Second Debater:
Of course not—if the system were unfair. But that’s why we build fairness in. You keep assuming the worst-case scenario is inevitable. We assume improvement is possible. One side believes in reform. The other seems to believe society is doomed no matter what we do.

Negative Second Debater:
No—we believe in learning from history. Facial recognition failed. Risk assessment tools failed. Predictive policing in Chicago failed. And each time, your answer is “try again with better rules.” How many failed experiments on marginalized communities before you admit the model itself is flawed?

Affirmative First Debater:
So every tool that starts poorly must be abandoned? By that logic, democracy should’ve been scrapped after the first corrupt election. Progress isn’t perfection—it’s course correction. And if we stop innovating because mistakes happen, then ethics becomes an excuse for stagnation.

Negative First Debater:
But when the mistake is someone being surveilled for life because their neighborhood has a high score, the cost isn’t just inefficiency—it’s lost trust, broken families, normalized suspicion. You treat error as a rounding issue. We treat it as a violation of human worth.

Affirmative Second Debater:
Preventive patrols aren’t experiments—they’re resource allocation. No arrests, no charges, just smarter deployment. You’re conflating presence with punishment. A cop walking down a street doesn’t mean everyone there is suspected. Light deters crime. So does intelligence.

Negative First Debater:
But when that light shines only on certain streets, generation after generation, it sends a message: You are dangerous. That’s not deterrence—that’s dehumanization. And when kids grow up seeing police arrive before anything happens, they don’t feel protected. They feel predicted.

This exchange captures the essence of free debate: rapid-fire logic, emotional resonance, and relentless pressure on core assumptions. The Affirmative frames AI as a tool of rational reform, insisting that abandoning it means accepting preventable harm. The Negative counters with moral urgency, arguing that some tools—no matter how efficient—corrode the very justice they claim to serve.

Both sides wield humor subtly: the Affirmative mocks resistance as Luddism; the Negative compares predictive policing to medical malpractice trials. Yet beneath the wit lies gravity—the recognition that this debate isn't about machines, but about what kind of society we choose to build.

Closing Statement

The closing statement is where a debate reaches its moral and intellectual crescendo. It is not merely a summary—it is a final act of persuasion, a synthesis of logic and value, designed to leave judges and audiences with one unshakable truth: which side has better defended what it means to live in a just society. On the motion “Is it ethical to use AI in predictive policing?”, both teams have grappled with questions far beyond technology—with the soul of justice itself. Now, in these final moments, they must answer not just what AI does, but what kind of world we want to build.

Affirmative Closing Statement

Ladies and gentlemen, let’s return to first principles.

Ethics isn’t about fear. It’s about consequences. It’s about choices. And our choice today is clear: do we allow preventable crimes to occur because we’re afraid of tools—or do we shape those tools to serve justice?

We’ve heard dramatic comparisons to Minority Report, dystopian visions of algorithmic tyranny. But let’s be honest: the real dystopia is a city where children grow up terrified of gunfire, where families lose loved ones to violence that could have been stopped. That’s the status quo—the world without intervention, without innovation.

Our opponents ask us to reject AI because it might be misused. But by that logic, we should ban cars because some drive recklessly, or outlaw medicine because pills can be abused. The ethical response isn’t prohibition—it’s regulation, oversight, and continuous improvement.

Let’s remember what we’ve proven:

First, AI saves lives—not through mind reading, but through intelligent resource allocation. When police are directed to high-risk areas before crime spikes, robberies drop. Assaults decline. Communities breathe easier. This isn’t speculation—it’s data from cities that dared to innovate responsibly.

Second, AI can reduce bias—when we demand transparency and fairness. Yes, bad models exist. So do biased officers. But unlike humans, algorithms can be audited, updated, and held accountable. A flawed algorithm can be fixed. A prejudiced gut feeling? That’s harder to patch. We don’t reject policing because officers make mistakes—we train, supervise, and reform. Why treat AI differently?

And third, this technology enables unprecedented accountability. Every decision traceable. Every input reviewable. With open-source models, civilian oversight boards, and mandatory impact assessments, we can create a system more transparent than any in history. Not perfect—but progressing.

The negative team says, “What if it goes wrong?” We say, “What if we never try?”

Because behind every statistic is a person. A mother who didn’t lose her son to a shooting. A shop owner who wasn’t robbed. A teenager diverted from gang recruitment because social services intervened early—not because they were suspected, but because they were supported.

We don’t live in a perfect world. But we can build a better one. One where technology doesn’t replace judgment—but sharpens it. Where data doesn’t erase dignity—but defends it.

So let us not fear the future. Let us guide it.

Because if ethics means doing the most good for the most people, then yes—using AI in predictive policing isn’t just ethical.

It’s essential.

Negative Closing Statement

Thank you.

Imagine being followed not because you did anything wrong—but because a machine decided your neighborhood was “risky.” Imagine your child being flagged by a system trained on decades of racist policing. Imagine living under constant surveillance, not as a suspect, but as a statistical likelihood.

That is not public safety.

That is pre-crime.

And today, we’ve shown why it cannot be ethical.

The affirmative team speaks of “saving lives” and “reducing bias,” but they offer dreams wrapped in code. They promise oversight, transparency, fairness—but where is it? Not in Los Angeles. Not in Chicago. Not in any major city using these systems. Instead, we find black boxes built by corporations, fed on poisoned data, deployed without consent.

Let’s be clear: predictive policing doesn’t predict crime. It predicts policing. It takes broken data—arrest records shaped by racism, poverty, and over-surveillance—and turns them into self-fulfilling prophecies. More patrols → more arrests → higher risk scores → even more patrols. The loop tightens. The community suffers. And the algorithm calls it “success.”

They say AI reduces human bias. But how? By replacing subjective prejudice with systemic bias encoded in mathematics. You can’t audit away the legacy of redlining. You can’t regulate away the fact that Black Americans are arrested at higher rates for the same behaviors. When AI learns from that data, it doesn’t correct history—it codifies it.

And let’s talk about their so-called “human-in-command” model. In theory, officers make the final call. In practice? Psychology shows we defer to machines, especially when they wear the mask of objectivity. An officer sees a red zone on a dashboard and thinks, “The computer knows best.” That’s not shared decision-making. That’s automation by suggestion.

They compare AI to vaccines and smoke detectors. But prevention requires proportionality. We don’t quarantine entire neighborhoods during a flu outbreak. We don’t wiretap every home to stop arson. So why accept blanket surveillance based on zip codes?

This isn’t about efficiency. It’s about power.

Who controls the algorithms? Private firms. Who bears the cost? Marginalized communities. Who benefits? Status quo enforcement.

History repeats itself: every time we automate injustice under the guise of neutrality, the vulnerable pay the price. From eugenics to stop-and-frisk, we’ve dressed oppression in scientific language before. Now we’re doing it again—with bigger datasets and fancier math.

We’re told, “Don’t throw out the scalpel.” But this scalpel cuts only one way. It doesn’t heal. It harms.

Because true justice isn’t about predicting risk—it’s about respecting rights. It’s about presumption of innocence. It’s about trusting people until they prove otherwise, not suspecting them because of where they live or how they look.

If we allow algorithms to decide who deserves scrutiny before any crime is committed, we haven’t advanced justice.

We’ve abolished it.

So let us choose differently.

Let us invest not in surveillance, but in schools. Not in prediction, but in opportunity. Not in control, but in care.

Because no algorithm should ever stand between a person and their freedom.

And no society that values justice can afford to forget that.