The specter of fatal violence haunts the corridors of human history like a relentless shadow, its contours shifting with each era’s technological and ideological tides. Yet, in the 21st century, a new sentinel emerges—not clad in armor or wielding a sword, but humming in the cold precision of silicon and code. Artificial intelligence, that most paradoxical of human creations, now stands poised to predict and prevent the very violence that has long defined our collective nightmares. But can a system forged in patriarchal logic truly dismantle the structures of oppression it was designed to perpetuate? The answer lies not in the algorithms themselves, but in the hands that wield them—and the feminist imagination that must guide their evolution.
The Illusion of Neutrality: AI as a Mirror of Human Bias
We cling to the myth of objectivity as if it were a sacred text, whispering that machines, devoid of emotion, must be free from the taint of human prejudice. Yet every dataset is a fossil record of historical inequities, and every predictive model is a palimpsest of the power structures that birthed it. When AI systems analyze patterns of violence, they do not see injustice—they see data points, correlations, and probabilities. A Black woman in an urban neighborhood is not a victim of systemic neglect; she is an “outlier” in a model trained on the screams of marginalized communities. The illusion of neutrality is the most insidious form of violence, because it renders oppression invisible while claiming to expose it.
Consider the chilling case of predictive policing algorithms, which, in their quest to “optimize” safety, merely amplify the biases of the justice systems they emulate. These systems do not predict violence—they predict which communities will be policed more aggressively, which bodies will be surveilled, and which lives will be deemed expendable. Feminist scholars have long argued that violence is not an aberration but a feature of patriarchal capitalism. If AI is to disrupt this cycle, it must first acknowledge that its “neutral” gaze is, in fact, a weaponized one—one that must be dismantled before it can be repurposed.
The Data Delusion: Whose Stories Get Saved?
Predictive AI thrives on data, and data is never neutral—it is curated, selected, and weaponized by those in power. The stories of women, particularly those from racialized or economically disenfranchised backgrounds, are often erased from the datasets that shape AI’s understanding of violence. A Latina domestic worker’s plea for help may never enter the training corpus of a domestic violence prediction model, while the arrest records of a wealthy white man accused of assault are meticulously logged. The result? A system that “predicts” violence based on who society deems worthy of protection—and who it deems disposable.
This is not merely an oversight; it is a structural erasure. Feminist data justice demands that we interrogate not just the outputs of AI, but the inputs—the silences, the absences, the stories that are deemed unworthy of being heard. If AI is to predict and prevent fatal violence, it must first learn to listen to the voices that have been systematically silenced. This requires a radical reimagining of data collection: one that centers the lived experiences of those most at risk, rather than the convenience of those in power. The alternative is a system that perpetuates the very violence it claims to combat.
The Ethics of Prediction: Can AI Truly Intervene?
Even if we could purge AI of its biases, the question remains: Can prediction alone prevent violence? The answer is a resounding no—unless intervention is rooted in community-led, feminist frameworks. AI can flag a high-risk individual, but it cannot dismantle the economic precarity that drives intimate partner violence. It can identify a pattern of online harassment, but it cannot heal the misogynistic culture that normalizes it. Prevention requires more than data—it requires solidarity, resources, and a commitment to dismantling the systems that produce violence in the first place.
Here, the limits of AI become glaringly apparent. Predictive systems operate on the logic of risk assessment, but risk is not an objective truth—it is a construct shaped by power. A Black trans woman may be flagged as a “high-risk” individual not because she is inherently dangerous, but because she exists at the intersection of multiple oppressions. The ethical dilemma is clear: Do we deploy AI to “protect” her by further surveilling her, or do we dismantle the structures that render her vulnerable in the first place? The latter requires a radical rethinking of safety—not as a top-down imposition, but as a collective, feminist project.
The Feminist Imagination: Reclaiming AI for Liberation
If AI is to be a tool of liberation rather than oppression, it must be reclaimed by feminist movements—not as a passive instrument, but as an active participant in the struggle for justice. This means centering the voices of those most affected by violence in the design and deployment of these systems. It means rejecting the myth of technological neutrality and embracing the messy, contradictory, and deeply human work of building alternatives. Feminist AI is not about optimizing violence prevention—it is about dismantling the conditions that make violence inevitable.
Imagine an AI system that does not predict violence but instead maps the pathways to liberation. It could identify gaps in social services, highlight the economic disparities that fuel domestic abuse, and amplify the demands of grassroots feminist organizations. It could be a tool for accountability, not surveillance—a way to hold institutions responsible for the violence they enable. But this requires a fundamental shift: from seeing AI as a savior to seeing it as a collaborator in a larger struggle for justice.
The Limits of Prevention: When AI Fails the Most Vulnerable
Even the most sophisticated AI systems have limits, and those limits are often the lives of the most marginalized. A trans woman of color may slip through the cracks of a predictive model not because the system is flawed, but because the system was never designed to see her. A disabled survivor of intimate partner violence may find that her experiences are too “atypical” for an algorithm trained on able-bodied narratives. The result is a system that fails those who need it most—a cruel irony in a world that claims to prioritize safety.
This is not a technical problem with a technical solution. It is a political problem that demands political answers. Feminist AI must grapple with the fact that prevention is not enough. We must also ask: Who is being protected? Who is being policed? And who is being erased? The answers to these questions will determine whether AI becomes a tool of liberation or a weapon of oppression.
The Future We Deserve: Beyond Prediction, Toward Transformation
The promise of AI to predict and prevent fatal violence is seductive—a sleek, futuristic solution to an ancient problem. But seduction is the language of patriarchy, and we must resist its allure. The future of AI is not predetermined; it is a battleground. Will we allow these systems to reinforce the very structures that oppress us, or will we seize them as tools for collective liberation? The choice is ours, but it is not a choice between “good” and “bad” AI—it is a choice between a world where violence is managed and a world where it is dismantled.
Feminist AI is not about building better mousetraps. It is about burning the house down.
