In the affluent community where I live, I walk down the street wearing a t-shirt of the elite Jesuit institution I once attended. It’s as if I presumptuously assume that the t-shirt will alleviate the prospect of being seen by surveilling gazes. Yet, I am instantaneously reminded as car doors lock when I walk beside them, and as purses clutch when I pass by, that my body is rendered as a site of suspicion. I am reminded that my Blackness precedes my existence. I am reminded that my darkened body occludes the presumption of innocence. I am reminded that my dark skin has eviscerated my humanity and has turned me into an object of fear. The embodied effects of being surveilled in this way leave me with fear, where my body and dignity have been distorted. As an object of suspicion, through the surveilling gaze interpreted through historically embedded racial meanings, the fear goes through me, as if I am different, unbelonging.

The preceding passage is a personal story about surveillance. Surveillance is, unfortunately, an inescapable reality of Black life. From a young age, I have lived with the knowledge that, despite how I see myself, this world has already told me who I am, who I should be, and that I should live with it. The containment, confiscation, and capture of my Black body through surveillance, a lived reality, is nonetheless my inheritance of centuries of stereotypes. Over time, these stereotypes have been reinforced within societal institutions and structures, creating insidious forms of systemic violence. With the emergence of digital algorithms and machine learning systems, these stereotypes are inherited by technology and, subsequently, codified into the “data and automated systems that now mediate our lives” (Hoffmann, 2018, para. 19). Surveillance, then, should not only be seen as a product of modern artificial intelligence (AI) technologies, such as automated facial recognition, drones, and predictive algorithms, but as part of a longstanding system of racialized hierarchies of control and dehumanization (Browne, 2015).

Sankofa: The Historical Genealogy of Racial Surveillance

AI is a black mirror that reflects our past, our worst impulses, moral failings, and the history of oppressive social, cultural, and institutional relationships. There is wisdom in learning from the past if we are wise enough to heed it. Beyond narratives of progress, innovative technologies, such as AI, cannot be separated from their historical context (Hampton, 2021). The concept of Sankofa, a word in the Akan language of Ghana that translates as “go back and fetch it,” reminds us to look back to see our way forward. As an invitation, Sankofa asks us to consider how contemporary AI technologies inherit and simultaneously reproduce longstanding traditions of systemic racism and racial injustice.

In Dark Matters: On the Surveillance of Blackness, Simone Browne (2015) demonstrates that contemporary forms of algorithmic surveillance are undergirded by a long history of racialized control and anti-Blackness. The scholar Akua Benjamin (2003) defines anti-Blackness as a distinct form of discrimination against individuals of Black African descent, rooted in their historical experiences of enslavement and colonization. Alongside this definition, anti-Blackness can also be understood as an antagonistic relationship between Blackness and the possibility of humanity (Dumas & Ross, 2016), sustained in what Saidiya Hartman (2007) calls the afterlife of slavery. In this afterlife of slavery, the potential harms of AI-powered surveillance should not be understood merely as glitches, technological errors, or accidental malfunctions, but as expressions of an entrenched “racially unequal past” (Mayson, 2019, p. 2297) in which new modes of racialized surveillance practices emerge.

Technological Capture: The Technological Afterlife of Racial Surveillance

According to Merriam-Webster, data is defined as factual information (such as measurements or statistics) used as a basis for reasoning, discussion, or calculation. Yet data is not always apodictic; it can be flawed. For instance, datasets used to train AI models may include exclusions and limitations in their collection practices and composition (Gebru et al., 2021). In particular, decades of unlawful, biased, and corrupt policing practices with disparate racial impacts have generated what is known as dirty data (Richardson et al., 2019). Consequently, AI-powered technologies developed in the name of justice that are modeled on dirty data—such as predictive analytics and surveillance systems—inherit and reproduce biases that govern how certain groups are classified, assessed, and perceived, creating a feedback loop of injustice.

Dirty dataisn’t a new phenomenon. But with the advent of AI, new light is being shed on an old problem (Mayson, 2019), in which historical patterns of racial bias are given new authority and reproduced through automated institutional decision-making. In view of this ongoing cycle of harm, AI surveillance technology risks producing new forms of anti-Blackness in the present and future, through a phenomenon I define as “technological capture”.

Technological capture describes how historical anti-Blackness is entangled with new technologies, such as AI, reconstituting conditions of capture in new computational forms. In other words, technological capture names not only the presence of racial bias within algorithmic systems, but also the historical and institutional process through which racialized social and institutional relationships become data, how that data becomes computational systems of classification, and how those computational classifications acquire material consequences when institutions act upon them.

Technological capture is not enacted by data, algorithms, or code alone. In what I call the ‘technological afterlife of racial surveillance’ it is not merely a technical problem. It is a human one. Capture is a lived racial condition enacted upon Black life: a form of containment, constraint, and disposability. Within this afterlife, AI-powered surveillance becomes consequential as inherited racial classifications are operationalized to monitor, predict, and reify boundaries along racial lines. Notably, the material consequences of technological capture depend on the institutions and actors that design, finance, deploy, and authorize these systems, including police officers, judges, administrators, agencies, and governing bodies that give algorithmic classifications material force. Here, new modes of racial profiling emerge in technological forms. Although technology changes the mechanism of capture, it has not necessarily ended the condition of capture itself.

The potential harm and consequences of technological capture are, unfortunately, evident in an increasing number of empirical examples of AI-powered surveillance technologies deployed by the criminal justice system. For example, a 2026 study found that networked surveillance technologies, such as Flock, were disproportionately concentrated in predominantly Black communities. A ProPublica investigation found that COMPAS, an algorithmic software used to assess recidivism risk, disproportionately produced false-positive high-risk classifications for Black defendants. Empirical research has also shown drastic disparities in error rates and misclassification between lighter- and darker-skinned individuals within automated facial recognition systems. Significantly, the quotidian nature of technological capture and its embodied effects are demonstrated by the wrongful arrest of Robert Williams, a Black man who was falsely identified by facial-recognition technology. As a violation of his human dignity, the wrongful arrest and detention of Robert Williams demonstrate that technological capture is more than a technical error when activated by the coercive power of the state.

A Faith That Does Justice: Healing, Accompaniment, Accountability, and Regulation

Efforts to mitigate, or abolish, technological capture require more than good intentions. Ethical and moral concern has sought to dismantle a long history of anti-Blackness emerging in and through contemporary technologies, such as AI, but has not rid the underlying logics of capture. Naming technological capture, therefore, not only exposes the persistence of anti-Blackness in emerging technologies but also demands the need to confront structural injustices and to center the communities who continue to be negatively impacted by systems of surveillance, control, and capture. With this in mind, the question should shift from how these systems cause harm to what justice demands in response. This question, especially as the potential harms of AI threaten the inviolable dignity of Black life, is intimately intertwined with the Jesuit tradition of a commitment of a faith that does justice.

In 1973, Fr. Pedro Arrupe gave a speech at the 10th International Congress of Jesuit Alumni of Europe in Valencia, Spain, in which he reflected critically on the need to reorient Jesuit education towards the promotion of justice and the liberation of the oppressed. In service to faith that does justice, Arrupe’s seminal call urges us to develop a moral imagination to conceive policies and social structures in service to the common good. To defend human dignity in the age of AI, Jesuit accompaniment must continue to heed Fr. Arrupe’s reorientation with creative fidelity, not only by standing in solidarity with the marginalized but also by taking appropriate action and advocacy.

As such, an adequate response -in the Jesuit tradition- to technological capture should be centered on four commitments: healing, accompaniment, accountability, and regulation.

Healing. This response begins first with historical memory by recognizing the racial histories inherited by AI-enabled surveillance systems (while also confronting the Society of Jesus’s own historical entanglement with slavery) and the demands of reconciliation that follow from it. In thisafterlife of slavery, the legacies of this violent past endure. Therefore, conversations about justice and technological advancement must begin first with remembering and healing this history.

Accompaniment. The second commitment is accompaniment, acknowledging and centering the lived experiences, knowledge, and agency of the communities most negatively affected by these new technologies. Accompaniment, in this sense, does not treat Black communities as recipients of justice but ensures they have a meaningful role and a seat at the table in matters of power, such as who benefits, who is harmed, and who hold authority. As an act of agency, accompaniment involves walking with these communities to determine how, where, and whether AI surveillance technologies are deployed. By centering those who are most affected by the outcomes of AI, while prioritizing community impact and participatory discernment on the utilization of AI systems, a Jesuit response guards against the threat of asymmetry of power in the development of emerging technology.

Accountability. The third commitment demands institutional accountability. In a report by the AI Now Institute titledAlgorithmic Impact Assessments Report: A Practical Framework for Public Agency Accountability(Reisman et al., 2018), the authors argue that “the turn to automated decision-making and predictive systems must not prevent agencies from fulfilling their responsibility to protect basic democratic values, such as fairness, justice, and due process” (p. 5). Without meaningful accountability mechanisms, AI-powered surveillance systems will continue to reproduce and intensify the injustices made visible through technological capture. For this reason, effective regulation is critical to mitigating the harm and risks posed by AI-powered surveillance technology.

Regulation. As the fourth commitment, regulation also includes transparency in how data used to train AI is collected, processed, and deployed. Currently, most AI systems remain opaque (West et al., 2019). To mitigate unwanted biases in the data AI models are trained on, regulation should include “restrictions or prohibitions on the use of the historical data generated by unlawful and biased practices” (Richardson et al., 2019, p. 47).

Given that regulation alone may still be insufficient to mitigate AI's potential harms, our response -in the Jesuit tradition- requires not only advocacy for effective regulation and clear limits on surveillance technologies that violate human dignity, but also a gesture toward algorithmic reparation. Algorithmic reparation focuses on repair by examining the injustices that algorithms inherit and unmasking the representational harms inflicted on affected communities (Davis et al., 2021). Algorithmic reparation also moves away from technical solutions that fail to address the underlying structural inequalities AI continues to amplify, and asks: should this system even exist at all? Justice, in this sense, may require more than accountability and regulation, and may point toward a structural refusal of AI surveillance altogether.

In the Jesuit tradition of faith that does justice, a response to the harms posed by technological capture must possess the creative fidelity and moral imagination to accompany affected communities, repair and redress injustices, and refuse AI technologies that cannot be reconciled with the inviolable dignity of human life.

References

• Benjamin, L. A. (2003).The Black/Jamaican criminal: The making of ideology (Publication No.305258209) [Doctoral dissertation, University of Toronto]. ProQuest Dissertations & Theses Global.
• Benjamin, R. (2019).Race after technology: Abolitionist tools for the new Jim code. John Wiley & Sons.
• Browne, S. (2015). Dark matters: On the surveillance of blackness. Duke University Press Books.
• Davis, J. L., Williams, A., & Yang, M. W. (2021). Algorithmic reparation. Big Data & Society, 8(2). https://doi.org/10.1177/20539517211044808
• Dumas, M. J., & ross, K. M. (2016). “Be real Black for me”.Urban Education, 51(4), 415-442. https://doi.org/10.1177/0042085916628611
• Gebru, T., Morgenstern, J., Vecchione, B., Vaughan, J. W., Wallach, H., Daumé III, H., & Crawford, K. (2021). Datasheets for datasets.Communications of the ACM, 64(12), 86–92. https://doi.org/10.1145/3458723
• Hampton, L. M. (2021). Black feminist musings on algorithmic oppression.Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency, 1-1. https://doi.org/10.1145/3442188.3445929
• Hartman, S. (2007). Lose your mother: A journey along the Atlantic slave route. Macmillan + ORM.
• Hoffmann, A. L. (2018, April 30).Data violence and how bad engineering choices can damage society. Medium. https://medium.com/@annaeveryday/data-violence-and-how-bad-engineering-choices-can-damage-society-39e44150e1d4
• Mayson, S. G. (2019). Bias in, bias out.Yale Law Journal, 128, 2218–2300.
• Reisman, D., Schultz, J., Crawford, K., & Whittaker, M. (2018).Algorithmic impact assessments report: A practical framework for public agency accountability. AI Now Institute.
• Richardson, R., Schultz, J. M., & Crawford, K. (2019). Dirty data, bad predictions: How civil rights violations impact police data, predictive policing systems, and justice.New York University Law Review Online, 94, 15–55.
• West, S.M., Whittaker, M., & Crawford, K. (2019).Discriminating systems: gender, race and power in AI. AI Now Institute.

Author

Cover image credits: Bias and discrimination in artificial intelligence. Design: Visual composition created by the SJES Communications Office. Background images: Photographs of real people used under a Canva Pro licence.