Safeguarding Human Rights in the Age of Artificial Intelligence: Evaluating the Adequacy of Legal Frameworks in criminal justice

Journal of Education, Communication, and Digital Humanities -Vol.2, No.1, Sept. 2025 pg. 201 Safeguarding Human Rights in the Age of Artificial Intelligence: Evaluating the Adequacy of Legal Frameworks in criminal justice Zainab Buba Department of General Studies Katsina State Institute of Technology and Management (KSITM) Email; [email protected] Phone no; 07031014580 DOI : https://doi.org/10.5281/zenodo.17387365 ABSTRACT The integration of Artificial Intelligence (AI) into criminal justice systems is reshaping law enforcement, adjudication, and corrections. From predictive policing to algorithmic risk assessments and sentencing recommendations, AI tools promise increased efficiency and consistency. However, their deployment raises critical legal and ethical concerns. This paper examines the impact of AI through the lens of fairness, accountability, transparency, and human rights, focusing on issues such as algorithmic bias, model opacity, and the reinforcement of systemic inequalities. It assesses whether existing legal frameworks constitutional protections, data privacy laws, and international human rights standards offer adequate safeguards. Drawing on comparative insights from jurisdictions using AI in justice, the analysis underscores both opportunities and risks, and calls for robust regulation and ethical oversight. The paper concludes that without strong legal safeguards, AI risks undermining fundamental rights and public trust in the justice system. Keywords: Artificial Intelligence (AI), Criminal Justice, Algorithmic Bias, Human Rights, Legal Safeguards Journal of Education, Communication, and Digital Humanities -Vol.2, No.1, Sept. 2025 pg. 202 1. Introduction Artificial Intelligence (AI) has emerged as one of the defining technologies of the twenty-first century, reshaping economies, governance, and social life. Its integration into criminal justice systems represents both an unprecedented opportunity and a substantial challenge. Law enforcement agencies, courts, and correctional institutions are increasingly adopting AI-driven tools such as predictive policing software, risk assessment algorithms, facial recognition technologies, and automated sentencing systems (Bryson, 2020). Advocates argue that such innovations enhance efficiency, reduce costs, and provide greater consistency in decisionmaking. However, critics caution that AI adoption in criminal justice risks entrenching existing inequalities, undermining human rights, and eroding public trust in the rule of law (Ferguson, 2017; Angwin et al., 2016; Barocas & Selbst, 2016). The core dilemma lies in the tension between technological innovation and the protection of fundamental rights. AI systems, while often promoted as objective, are not immune to bias; numerous studies show that predictive policing and risk assessment tools disproportionately target marginalized communities, reproducing systemic discrimination embedded in historical data. These concerns are amplified in contexts where institutional safeguards are weak or unevenly enforced, raising urgent questions about fairness, transparency, and accountability. Recent legal and policy developments underscore the global dimensions of this debate. In the Global North, the European Union’s Artificial Intelligence Act (2024) has introduced the world’s first comprehensive regulatory framework, prohibiting high-risk practices such as predictive policing and mandating robust oversight of justice-related AI systems. In the Global South, Nigeria’s Data Protection Act (2023) establishes a national legal foundation for safeguarding digital rights, including limits on automated decision-making, while exposing persistent challenges of enforcement capacity and digital exclusion. African scholarship and policy dialogues, such as UNESCO’s 2025 East Africa forum on AI and the rule of law, further highlight the need for locally grounded approaches that avoid Eurocentric framings of justice and rights. This study situates the use of AI in criminal justice within these intersecting debates, drawing on both Global North and Global South perspectives. By examining how emerging legal frameworks, socio-technical realities, and human rights concerns converge, it aims to illuminate the promises and perils of deploying AI in justice systems, and to assess the extent to which regulatory models can balance innovation with equity and accountability. From a legal perspective, the deployment of AI in criminal justice intersects with constitutional guarantees, statutory protections, and international human rights frameworks. Central to this debate are questions of fairness, accountability, and transparency. Do existing legal frameworks sufficiently safeguard against algorithmic harms? If not, what reforms are necessary to ensure that AI serves the cause of justice rather than undermines it? Addressing these questions requires a careful examination of current laws, an analysis of case studies from different jurisdictions, and a forward-looking discussion of regulatory and ethical approaches. The significance of this inquiry cannot be overstated. Criminal justice systems wield immense power over individual lives, liberty, and dignity. Any technology integrated into such systems must therefore meet the highest standards of fairness and accountability. Unlike commercial applications of AI, where errors may result in financial loss or reputational harm, errors in criminal justice contexts may result in wrongful arrests, unjust convictions, or disproportionate Journal of Education, Communication, and Digital Humanities -Vol.2, No.1, Sept. 2025 pg. 203 sentencing. As such, the stakes of AI governance in this domain are uniquely high (Crawford, 2021). This paper proceeds in seven parts. Following this introduction, Part Two provides an overview of AI applications in criminal justice, highlighting both domestic and international examples. Part Three examines the potential opportunities of AI, particularly its contributions to efficiency, consistency, and crime prevention. Part Four explores the legal and ethical concerns surrounding AI use, with attention to bias, transparency, accountability, and due process. Part Five evaluates the adequacy of existing legal frameworks, analysing constitutional protections, data protection regimes, and international human rights obligations. Part Six proposes pathways toward stronger legal safeguards, emphasizing principles of fairness, accountability, and transparency, alongside regulatory and policy reforms. Finally, Part Seven concludes with reflections on balancing technological innovation with the imperatives of justice and human rights. By situating AI within the broader discourse of law, technology, and society, this paper contributes to an urgent and evolving debate. It seeks to illuminate not only the risks of unchecked AI adoption but also the opportunities for harnessing technology in service of a more equitable and effective criminal justice system. 2. Artificial Intelligence in Criminal Justice: An Overview Artificial Intelligence is increasingly woven into the fabric of modern criminal justice systems, where it is deployed across the spectrum of law enforcement, adjudication, and corrections. AI in this context generally refers to computer systems capable of analysing vast amounts of data, identifying patterns, and making predictions or recommendations that inform decision-making. While such systems promise efficiency and enhanced accuracy, their application in criminal justice is uniquely sensitive due to the high stakes involved—personal liberty, human dignity, and societal trust in legal institutions. 2.1 Predictive Policing Predictive policing is among the most prominent applications of AI in criminal justice. It involves using historical crime data and machine-learning algorithms to forecast where crimes are likely to occur or who might be at risk of committing or experiencing crime (Perry, McInnis, Price, Smith, & Hollywood, 2013). Police departments in cities such as Los Angeles, Chicago, and London have experimented with predictive policing tools like PredPol and HunchLab. While these programs claim to optimize resource allocation and reduce crime rates, empirical studies reveal mixed outcomes. In practice, predictive policing often reflects and amplifies existing biases in policing data, disproportionately targeting minority communities already subject to over-policing (Lum & Isaac, 2016). This dynamic not only perpetuates cycles of criminalization but also raises concerns about equal protection under the law and the right to non-discrimination. Journal of Education, Communication, and Digital Humanities -Vol.2, No.1, Sept. 2025 pg. 204 2.2 Risk Assessment Tools Another key application of AI is in risk assessment instruments used during pretrial, sentencing, and parole decisions. Tools such as the Correctional Offender Management Profiling for Alternative Sanctions (COMPAS) in the United States are designed to predict the likelihood of reoffending. Judges and parole boards rely on these assessments to inform decisions about bail, sentencing length, and parole eligibility. Although risk assessments are promoted as objective alternatives to human judgment, research indicates that they are susceptible to significant error and bias. A landmark investigation by ProPublica found that COMPAS disproportionately overestimated recidivism risk for Black defendants while underestimating it for White defendants (Angwin et al., 2016). Such outcomes undermine the fairness of sentencing and challenge the foundational legal principle of equality before the law. 2.3 Facial Recognition Technologies Facial recognition software represents another controversial application of AI in criminal justice. Law enforcement agencies use these systems for surveillance, suspect identification, and forensic investigations. Countries such as China have adopted facial recognition extensively for public security purposes, while police forces in the United States and Europe have deployed it in varying degrees. However, numerous studies highlight the inaccuracy of facial recognition, particularly in identifying women, people of colour, and younger individuals (Buolamwini & Gebru, 2018). Misidentification can lead to wrongful arrests and prosecutions, posing significant risks to due process and personal liberty. Furthermore, widespread surveillance facilitated by facial recognition technology raises pressing questions about privacy rights and proportionality in democratic societies. 2.4 Automated Sentencing and Decision Support Some jurisdictions are exploring AI systems to provide decision-support tools for judges. These tools may recommend sentencing ranges, assess flight risk, or suggest alternatives to incarceration. Proponents argue that automation can reduce disparities caused by human bias, while critics contend that embedding algorithms into sentencing may entrench rather than mitigate inequalities (Binns, 2019). Moreover, reliance on automated recommendations challenges judicial independence and discretion. If judges defer excessively to algorithmic outputs, accountability for sentencing decisions may become obscured. This undermines the principle that justice must not only be done but must also be seen to be done. 2.5 Comparative perspectives The deployment of AI in criminal justice varies significantly across jurisdictions. In the United States, adoption is largely decentralized, with local police departments and state courts experimenting with different tools. In contrast, the European Union has adopted a more cautious approach, emphasizing fundamental rights and proposing stringent regulations under the draft Artificial Intelligence Act (European Commission, 2021). China, by comparison, has Journal of Education, Communication, and Digital Humanities -Vol.2, No.1, Sept. 2025 pg. 205 embraced AI widely in law enforcement and judicial processes, reflecting its broader model of governance and surveillance. These divergent approaches illustrate the global patchwork of AI governance. While some jurisdictions prioritize innovation and efficiency, others stress human rights protections. The comparative analysis underscores the urgent need for international dialogue and cooperation to develop common principles that reconcile technological advancement with the imperatives of justice. In sum, AI is already embedded in various facets of criminal justice, from predictive policing to facial recognition and risk assessment. Each application carries potential benefits but also significant risks, particularly in relation to fairness, accountability, and human rights. Understanding these technologies and their real-world consequences is essential for evaluating their compatibility with the rule of law and for developing appropriate regulatory response. Opportunities of AI in Criminal Justice The adoption of Artificial Intelligence (AI) in criminal justice is not without justification. Proponents emphasize that AI-driven tools have the potential to enhance efficiency, improve consistency, and enable data-driven strategies that can transform how justice systems operate. These opportunities, if harnessed responsibly, could contribute to fairer and more effective outcomes in law enforcement, adjudication, and corrections. 3.1 Efficiency and Resource Optimization Criminal justice systems worldwide often struggle with resource constraints, ranging from understaffed police departments to overburdened courts. AI tools offer a means of optimizing limited resources by automating routine tasks and enabling faster decision-making. Predictive policing, for instance, allows law enforcement agencies to allocate personnel more strategically, concentrating efforts in areas where crimes are statistically more likely to occur (Perry et al., 2013). Similarly, automated document review systems can assist prosecutors and defence attorneys by rapidly analysing large volumes of evidence, reducing delays in trials. In correctional facilities, AI-driven monitoring systems can enhance security while reducing labour costs. These efficiencies do not merely represent cost savings; they may also translate into shorter case backlogs, faster adjudication, and improved access to justice for individuals awaiting trial. 3.2 Enhancing Consistency and Objectivity One of the most compelling arguments in favor of AI adoption is its potential to reduce human bias and inconsistency. Human decision-makers in criminal justice—police officers, judges, parole boards—are susceptible to subjective biases, fatigue, and error. AI systems, by contrast, apply standardized algorithms to similar cases, theoretically ensuring greater consistency across decisions (Bryson, 2020). For example, sentencing algorithms may help establish uniformity by recommending penalties based on structured criteria rather than personal discretion. In jurisdictions where disparities in sentencing for similar crimes have eroded public trust, algorithmic tools could play a corrective role by curbing arbitrary variations. However, in Nigeria, the introduction of AI into criminal justice raises distinctive risks that differ from the Western contexts in which most studies have been conducted. Predictive policing, for instance, could entrench existing policing patterns that already disproportionately target certain ethnic or regional groups. Nigeria’s history of security operations in the Niger Journal of Education, Communication, and Digital Humanities -Vol.2, No.1, Sept. 2025 pg. 206 Delta, the North-East, and conflict-prone Middle Belt illustrates how policing is often interwoven with ethnic and political tensions (Akinola, 2020; HRW, 2020). If predictive algorithms are trained on historical arrest and crime data, they are likely to replicate and amplify these biases, reinforcing ethnic profiling and deepening mistrust between communities and law enforcement. Moreover, Nigeria’s criminal justice institutions face persistent challenges of weak oversight, limited transparency, and inconsistent enforcement of rights protections. Judicial review mechanisms are often slow, under-resourced, or inaccessible to marginalized defendants (Okagbue, 2021). In such a context, the opacity of AI “black box” systems could further erode accountability, as defendants may lack effective avenues to contest algorithmic decisions. This risk is compounded by gaps in the enforcement capacity of the Nigeria Data Protection Commission, which is still developing the expertise and infrastructure necessary to audit AIdriven systems under the Data Protection Act (Federal Republic of Nigeria, 2023; ThisDayLive, 2025). Thus, while algorithmic tools hold promise for promoting consistency, their deployment in Nigeria could reinforce rather than correct systemic inequities if not carefully regulated. Far from eliminating bias, predictive policing and risk assessment systems may harden pre-existing divisions, exacerbating inter-ethnic tensions and weakening already fragile trust in the rule of law. 3.3 Data-Driven Crime Prevention The predictive capacity of AI is particularly valuable for crime prevention. By analysing large datasets—including crime statistics, socioeconomic indicators, and spatial patterns—AI can identify trends that inform proactive interventions. Predictive policing, when carefully regulated, may allow authorities to address emerging crime “hotspots” before incidents occur, thereby improving public safety (Lum & Isaac, 2016). Additionally, AI can enhance investigative capabilities by identifying links in complex data sets, such as connections between financial transactions in money-laundering cases or communications in organized crime networks. This ability to uncover hidden patterns may empower law enforcement to dismantle criminal enterprises more effectively. Yet, the risks of predictive policing in Nigeria diverge significantly from those documented in Western contexts. Historical patterns of policing in Nigeria reveal systemic ethnic profiling and regional disparities in law enforcement. Security operations in the Niger Delta, counterterrorism campaigns in the North-East, and farmer–herder conflicts in the Middle Belt have long been shaped by ethnic and political fault lines (Akinola, 2020; HRW, 2020). If predictive algorithms are trained on such datasets, they could disproportionately flag HausaFulani communities in the North, Ijaw and Ogoni groups in the Delta, or Tiv farmers in the Middle Belt as “high risk.” In a society where policing already suffers from accusations of bias and selective enforcement, algorithmic policing may harden ethnic stereotypes, deepen intercommunal grievances, and risk inflaming tensions in already volatile regions. These dangers are magnified by Nigeria’s policing practices and weak judicial oversight. The Nigeria Police Force has triggered the nationwide #EndSARS protests (Amnesty International, 2020). Embedding such practices into algorithmic systems risks automating discrimination at scale, lending a veneer of “scientific objectivity” to biased outcomes. Furthermore, Nigeria’s judiciary lacks the capacity to provide timely oversight. Chronic delays, underfunding, and Journal of Education, Communication, and Digital Humanities -Vol.2, No.1, Sept. 2025 pg. 207 restricted access to legal remedies mean that defendants may find it nearly impossible to challenge algorithmically generated “risk scores” or predictive policing decisions (Okagbue, 2021). The opacity of AI “black box” systems further compounds this problem, as neither defendants nor courts are likely to access the logic driving such outputs. 3.4 Supporting Judicial Decision-Making AI can serve as a decision-support tool for judges and legal practitioners. Risk assessment systems, for example, provide data-driven evaluations of a defendant’s likelihood of reoffending, which may assist in bail and parole determinations. While such systems are not free from controversy, they can supplement judicial reasoning by offering perspectives grounded in statistical analysis (Barocas & Selbst, 2016). In the United States, however, the COMPAS tool has been criticized for racially biased outcomes, raising constitutional concerns about due process in State v. Loomis (2016). These global experiences highlight the risks of uncritical adoption and provide important lessons for Nigeria. In Nigeria, the judicial context presents distinctive challenges. Chronic case backlogs, underfunding, and shortages of trained personnel often pressure judges to seek efficiency at the expense of deliberation (Okagbue, 2021). Against this backdrop, there is a real danger that algorithmic outputs may be treated as authoritative, reducing judges to “rubber stamps” for opaque technologies. This could erode the constitutional right to fair hearing under Section 36 of the 1999 Constitution, especially where defendants lack the resources to contest adverse scores. Nigeria’s justice system is also shaped by longstanding ethnic, socio-economic, and regional disparities. Patterns of harsher pre-trial detention for young men from marginalized communities, as documented during the #EndSARS protests, risk being amplified if encoded into algorithmic models (Amnesty International, 2020; Akinola, 2020). An algorithm trained on such data may systematically classify certain groups—such as urban youth in Lagos or minority populations in conflict-prone areas—as “high risk,” reinforcing discriminatory practices under the guise of objectivity. Beyond risk assessments, natural language processing (NLP) tools could improve judicial efficiency by helping judges, lawyers, and self-represented litigants navigate statutes and precedents. In principle, this could reduce information asymmetries in a system where many defendants lack adequate legal representation. However, Nigeria’s uneven digital infrastructure, limited ICT capacity in many courts, and persistent digital exclusion of rural populations raise the risk that AI-enabled legal research may widen, rather than bridge, gaps in access to justice (UNESCO, 2025). These challenges suggest that AI in judicial decision-making is not simply a technical question but a governance issue. Without safeguards, the technology could undermine, rather than strengthen, Nigeria’s justice system. Effective adoption requires: 1. Mandatory transparency standards, including disclosure of algorithmic logic in judicial contexts. 2. Bias and impact assessments before deploying AI in bail, parole, or sentencing decisions. 3. Judicial training to ensure judges understand AI outputs as advisory, not determinative. Journal of Education, Communication, and Digital Humanities -Vol.2, No.1, Sept. 2025 pg. 208 4. Investment in digital infrastructure to avoid deepening inequalities in access to justice. Thus, while AI-enabled decision-support tools hold promise, Nigeria’s institutional weaknesses mean that their unregulated use could entrench systemic bias and undermine constitutional guarantees. Careful design, strong oversight, and alignment with human rights norms are essential if AI is to serve as a tool for justice rather than a new layer of inequality. 3.5 Advancing Transparency and Accountability (Potentially) Although often criticized for their opacity, AI systems also hold the potential to advance transparency in some contexts. When designed with explain ability features, algorithms can document the reasoning process behind decisions more consistently than humans, whose motivations may be opaque or unrecorded. This capacity for systematic documentation could facilitate review and oversight, provided that algorithms are subject to rigorous auditing (Binns, 2019). For example, automated sentencing tools that log their decision-making criteria could create auditable trails, making it easier to detect inconsistencies or biases compared to purely human judgments. Thus, under the right conditions, AI may strengthen rather than weaken accountability. 3.6 Comparative Benefits across Jurisdictions Different legal systems stand to benefit from AI in distinct ways. In developed jurisdictions, AI may primarily enhance efficiency and uniformity in systems already flush with resources. In developing contexts, where criminal justice institutions may suffer from chronic underfunding and case backlogs, AI could provide transformative gains in access to justice by reducing workload burdens and supporting overstretched personnel (Crawford, 2021). Furthermore, cross-border applications of AI, such as international cooperation in combating cybercrime or terrorism, illustrate its potential to augment global security efforts. By facilitating collaboration across jurisdictions, AI can contribute to more coordinated responses to transnational threats. 3.7 Summary The opportunities presented by AI in criminal justice are significant. From optimizing resource allocation and promoting consistency to enhancing predictive capabilities and supporting judicial decisions, AI technologies offer the promise of a more efficient and equitable system. However, these benefits are not automatic. They are contingent upon careful design, rigorous oversight, and alignment with legal and ethical standards. Without such safeguards, the very advantages AI purports to deliver may be undermined by unintended harms. 4. Legal and Ethical Concerns While Artificial Intelligence (AI) offers compelling opportunities for improving efficiency and consistency in criminal justice systems, it also raises profound legal and ethical challenges. Because criminal justice involves decisions that directly affect individual liberty and fundamental rights, the risks associated with AI adoption are particularly acute. These concerns Journal of Education, Communication, and Digital Humanities -Vol.2, No.1, Sept. 2025 pg. 209 can be grouped into five broad categories: algorithmic bias, transparency and explain ability, accountability and liability, due process and fair trial rights, and privacy and surveillance. 4.1 Algorithmic Bias and Discrimination One of the most pervasive criticisms of AI in criminal justice is its susceptibility to bias. AI systems learn from historical data, which often reflects entrenched patterns of discrimination in policing, sentencing, and corrections. As a result, algorithms may reproduce and even exacerbate systemic inequalities. The most cited example is the COMPAS risk assessment tool in the United States. A 2016 ProPublica investigation revealed that COMPAS systematically overestimated the likelihood of recidivism for Black defendants while underestimating risk for White defendants (Angwin et al., 2016). This disparity not only undermines fairness but also contravenes the principle of equal protection under the law. Bias has also been documented in predictive policing systems. By relying on historical crime data that disproportionately reflects arrests in minority communities, predictive policing directs law enforcement resources back to those same areas, creating a feedback loop of over-policing (Lum & Isaac, 2016). Such practices risk reinforcing racial and socioeconomic inequalities, raising concerns under constitutional and human rights law. Transparency is central to the legitimacy of legal systems, yet many AI tools function as “black boxes.” Their underlying algorithms are often proprietary, complex, and inaccessible, even to experts. This opacity poses a serious challenge in legal contexts where decisions must be reviewable and subject to challenge. For instance, when risk assessment tools influence bail or sentencing, defendants and their counsel may lack the ability to scrutinize how risk scores were calculated. Without access to the underlying logic, contesting these outcomes becomes nearly impossible, undermining due process rights (Goodman & Flaxman, 2017). The issue is further complicated by trade secrecy claims. Technology companies often refuse to disclose algorithmic details on the grounds of protecting intellectual property. Courts have been divided on whether such claims outweigh defendants’ rights to a fair trial. In State v. Loomis (2016), the Wisconsin Supreme Court allowed the use of COMPAS in sentencing, despite acknowledging its opacity, provided it was not the sole basis for the decision. This compromise illustrates the legal system’s struggle to balance innovation with transparency. 4.3 Accountability and Liability Accountability is a cornerstone of justice, yet AI challenges traditional frameworks of legal responsibility. When an algorithm produces a flawed recommendation—such as a miscalculated risk score leading to unjust sentencing—who should be held accountable? The judge who relied on the tool, the developer who designed it, or the institution that adopted it? Current legal systems are ill-equipped to address such questions. In many jurisdictions, accountability is diffused across multiple actors, creating a vacuum of responsibility. This lack of clarity risks eroding public trust, as victims of algorithmic errors may struggle to identify a responsible party. Journal of Education, Communication, and Digital Humanities -Vol.2, No.1, Sept. 2025 pg. 216 To address these challenges, Nigeria should pursue a comprehensive reform agenda that combines legal innovation, judicial vigilance, institutional strengthening, and international collaboration. The following policy recommendations summarize the core priorities identified in this study: • Legislative Reform: Amend the NDPA 2023 to explicitly regulate automated decision-making, mandate human oversight, and require transparency in the deployment of AI within criminal justice. • Judicial Oversight: Develop practice directions and evidentiary standards to ensure that AI technologies do not compromise the constitutional right to a fair hearing. • Privacy Protections: Restrict AI-driven surveillance to cases authorized by independent judicial approval, in line with constitutional guarantees of privacy. • Bias Prevention: Introduce mandatory algorithmic impact assessments and anti-discrimination audits prior to deploying AI in law enforcement or judicial processes. • Institutional Strengthening: Invest in the capacity of courts, regulators, and oversight bodies to evaluate AI systems and enforce compliance with rights-based standards. • Regional and International Cooperation: Engage actively in African Union initiatives and align national frameworks with global instruments such as the UN Guiding Principles on Business and Human Rights (2011), the OECD AI Principles (2019), and the UNESCO Recommendation on the Ethics of AI (2021). In conclusion, the challenge for Nigeria—and indeed for the global community is to strike a careful balance between harnessing the benefits of AI and safeguarding fundamental rights. Achieving this balance requires foresight, strong institutions, and alignment with international norms. By adopting proactive reforms, Nigeria can protect its citizens from the risks of AIdriven criminal justice while positioning itself as a leader in rights-based AI governance across Africa. References • Akinola, A. O. (2020). Ethnic profiling and policing in Nigeria: Implications for human rights and justice. African Security Review, 29(3), 273–290. • Amnesty International. (2020). Nigeria: Authorities must end impunity for police violence in wake of #EndSARS protests. Retrieved from https://www.amnesty.org • Angwin, J., Larson, J., Mattu, S., & Kirchner, L. (2016). Machine bias: There’s software used across the country to predict future criminals. And it’s biased against blacks. ProPublica. • Barocas, S., & Selbst, A. D. (2016). Big data’s disparate impact. California Law Review, 104(3), 671–732. • Bridges v. South Wales Police [2020] EWCA Civ 1058 (UK Court of Appeal). • Bryson, J. (2020). The past decade and future of AI’s impact on society. In M. C. Horowitz (Ed.), AI and International Affairs. Brookings Institution Press. • Crawford, K. (2021). The Atlas of AI: Power, Politics, and the Planetary Costs of Artificial Intelligence. Yale University Press. Journal of Education, Communication, and Digital Humanities -Vol.2, No.1, Sept. 2025 pg. 217 • Digital Rights Lawyers Initiative v. National Identity Management Commission (NIMC), Suit No. FHC/ABJ/CS/79/2020 (Federal High Court, Nigeria). • Emergent Markets Telecommunication Services Ltd. v. Eneye (2018) LPELR46189(CA) (Court of Appeal, Nigeria). • European Commission. (2024). The EU AI Act enters into force. Brussels: European Union. Retrieved from https://digitalstrategy.ec.europa.eu/en/policies/regulatory-framework-ai • European Union. (2021). Proposal for a Regulation laying down harmonised rules on Artificial Intelligence (Artificial Intelligence Act). COM/2021/206 final. • Federal Republic of Nigeria. (2023). Nigeria Data Protection Act (NDPA). Official Gazette, Abuja. • Ferguson, A. G. (2017). The rise of big data policing: Surveillance, race, and the future of law enforcement. NYU Press. • General Data Protection Regulation, Regulation (EU) 2016/679. • Human Rights Watch (HRW). (2020). Nigeria: Events of 2020. World Report 2021. Retrieved from https://www.hrw.org • Lum, K., & Isaac, W. (2016). To predict and serve? Significance, 13(5), 14–19. • OECD. (2019). OECD Principles on Artificial Intelligence. Paris: OECD. • Okafor v. Lagos State Government (2016) (High Court of Lagos State, unreported). • Okagbue, I. (2021). Justice sector reforms in Nigeria: Challenges and prospects. Journal of African Law, 65(1), 87–105. • State v. Loomis, 881 N.W.2d 749 (Wis. 2016) (Supreme Court of Wisconsin, USA). • ThisDayLive. (2025, May 3). Addressing data privacy concerns in artificial intelligence systems: Regulatory mechanisms in Nigeria. Retrieved from https://www.thisdaylive.com • Ubani v. Director, SSS (1999) 11 NWLR (Pt. 625) 129 (Court of Appeal, Nigeria). • UNESCO. (2021). Recommendation on the Ethics of Artificial Intelligence. Paris: UNESCO. • UNESCO. (2025, February 29). Harnessing AI for justice: Balancing innovation and equity in East Africa. Retrieved from https://www.unesco.org/en/articles/harnessing-ai-justice-balancinginnovation-and-equity-east-africa • United Nations. (2011). Guiding Principles on Business and Human Rights. Geneva: United Nations.