Colonial Judicial Legacy as a Latent Challenge for the Adoption of Algorithmic Sentencing in African Courts

[Jake Okechukwu Effoduh is an Assistant Professor at the Lincoln Alexander School of Law of Toronto Metropolitan University, where he teaches Critical Approaches to Data, Algorithm and Science in the Law. He has been an international human rights lawyer for fourteen years.]

Introduction

The pace at which Artificial Intelligence (AI) is transforming the face of almost every human activity cannot be ignored or overemphasized. The impacts of AI range from health, finance, telecommunication, agriculture, and security to education, administrative decision-making, etc. The judiciary is not exempted, with some countries experimenting with the use of AI algorithms in judicial decision-making, including criminal sentencing. Thus, some jurisdictions are shuttling between traditional human judges to AI algorithmic predictive justice. China may be topping the list of jurisdictions that have since embraced the use of AI in criminal trial procedures, including bail, probation, parole, and sentencing. AI tools like “Little Judge Bao” are used to provide sentencing recommendations and assist judges in making decisions based on big data and predictive algorithms. In the U.S., one such program, “COMPAS,” has been used in assistive judicial decision-making by some United States courts despite facing controversies.

Certain African legal systems are considering incorporating AI programs into their practice to enhance the effectiveness of their justice systems. Although still in development and not yet capable of automated court decisions, these programs are currently being deployed for complex legal information analysis and identification. Such tools are being experimented with within the legal practice of some West African countries like Nigeria and East African countries like Kenya, South Africa, and Uganda.

It then seems urgent to ask: What are fundamental challenges that could arise when African courts fully adopt AI tools for criminal justice decision-making? While benefits have been identified, such a transformative decision is bound to erupt inherent judicial challenges that are not directly bound to the AI tools themselves. These judicial problems, often downplayed or ignored but deeply rooted in the judicial history of most African countries, will obscure whatever benefits accruable from the adoption of AI tools by African courts. This article identifies these challenges and proposes strategies to maximize the deliverables of AI tools in African courts.

Fundamental and Peculiar Challenges to Adopting AI in African Courts

Ignoring the complex colonial judicial history of African countries while building AI systems for their courts is bound to be problematic. Technology transfer must consider the contextual realities and peculiarities of the society it is being transferred to for it to be practically beneficial. Some systematic scholarships have analyzed the impacts of colonialism on Africa. These studies reveal that colonialism touches on all aspects of the continent’s life, including the criminal justice system.

Colonial judicial legacies and their impacts have become embedded and remain indelible in Africa’s rules and practices regarding the administration of justice. These constitute potential risks to the adoption of AI tools in almost any African court. This article analyzes five prominent colonial judicial legacies which are considered potential challenges to adopting AI systems for criminal sentencing in African courts, viz: A Supplanted Purpose of Criminal Sentencing, Inherited Judicial Corruption, Entrenchment of Partiality, Conflict of Law and AI Programming, and Infrastructural Deficits.

(1) A Supplanted Purpose of Criminal Sentencing

Even before Africa was colonized, highly organized and hierarchal justice systems existed across various African societies. Sentencing, which constitutes the penultimate stage in most African criminal justice procedures, had a key purpose – the maintenance of social equilibrium. When a person was sentenced, the central aim was not the punishment itself but the keeping of societal harmony. In fact, to maintain societal sanity at all costs, family members of an offender could sometimes collectively suffer from the impact of a punishment apportioned to the offender. For instance, public opprobrium could dent the image of the offender’s family members. Thus, retribution had little or no place in pre-colonial African criminal sentencing; the focus was on restorative justice— maintaining peace between the disputing parties and the relevant community.

Pre-colonial African judges were trained to prioritize this approach before colonial justice sentencing principles, which prioritized the retributive principle of criminal sentencing, supplanting previous systems. The persistence of this inherited colonial sentencing culture in African legal systems has led to a protracted legal conflict that still requires resolution. Programing such incompatible legal heritage into AI tools for use by African courts in an age where decolonization of African law is a front-burner discourse will only serve to perpetuate and complicate matters further.

(2) Inherited Judicial Corruption

Whether among the Yoruba and Igbos of southern Nigeria, Masai of Kenya, Nguni of southern Africa, or the Nubians in northern Africa, judicial procedures were never monetized. Justice was strictly a right for those who deserved it, not for the affluent or highest bidders. There were no court fees, so a wronged person did not need to be wealthy to access the court and drink from the source of justice. More importantly, the African criminal justice system was hinged on certain religious beliefs. There was a strong faith in supernatural participation in the adjudication process which deterred judges from serving corrupt justice. For instance, among the people of southeastern Nigeria, the Arochukwu god “Ibini Ukpabi” was feared and respected as a true justice-dispensing god. Similar gods of justice, like Shango and Ogun, were worshipped by the Yoruba people of southwestern Nigeria. There existed the fear of earthly and nirvana punishments for those who violated societal justice principles and ethics. This spiritual and religious involvement of supernatural forces served as effective checks and balances on pre-colonial judges, thereby making judicial corruption an extremely feared sacrilege.

However, with the arrival of the colonial judicial system came the monetization of justice procedures, accompanied by deliberate satanization, denigration, and desecration of the African gods of justice. This engendered institutional corruption and eventually clouded the indigenous African notion of justice. Documented judicial corrupt practices are exemplified by what was perpetrated by colonial magistrate Benjamin Way and allegations against court clerk Augustus Otonba-Payne. Issues of court rules occurred under Benjamin Way’s oversight like fabrication, bribery, “selling” of justice, and conflict of interest activities. Although colonialism may have ended in principle, its culture of judicial corruption continues to influence most African courts’ practice rules. AI tools trained with such existing rules of practice are bound to be problematic.

(3) The Entrenchment of Partiality

Before encountering their colonizers, African societies were inhabited by people who, while exhibiting inter-group heterogeneity, shared significant intra-group similarities in culture, history, belief systems, language, cuisine, etcetera. These similarities forged a common bond among the people, making racial discrimination and other divisive tendencies that could inappropriately influence justice and criminal sentencing rare. The beat changed when the arrival of colonialism brought in people of entirely different races, traditions, interests, and mindsets of superiority. This led to artificial and prejudicial differentiation of parties before the court. Criminal trials expanded to include actions between Africans and Europeans and actions against the state represented by colonial authorities. Judicial prejudices based on colour and the notion of racial superiority often resulted in Africans rarely receiving justice in colonial courts.

Several unreported nineteenth-century cases analysed by Omoniyi Adewoye in his book “The Judicial Systems in Southern Nigeria 1854-1954” (pp. 78—79), like R. v. Lobley (1869) and R. v. Stott (1869), demonstrate how partiality became entrenched in the African criminal justice system through discriminatory judgments and sentencing practices. In both cases, an African and a European who committed similar criminal offences received different sentences; the European involved received four months’ imprisonment, while the African received three years imprisonment. This pattern of sentencing discrimination remains in most African courts, where the affluent and influential escape the sword of the law while commoners are hit hard by the hammer of justice. Even if AI is configured for impartial sentencing, colonial judicial legacy may find ways into the training data and models, which may lead to outcomes that could disproportionally impact the Africans with less power.

(4) Conflict of Laws and AI

ProgrammingIn pre-colonial times, African societies’ justice systems functioned through African customary law, which, although diverse and initially unwritten, regulated civil and criminal justice with well-established principles. This law was ascertainable through custodians, who maintained the sanctity of the law and served as oral repositories. However, colonialists condemned African customary law as barbaric, outdated, and inadequate to cope with realities. As such, European laws based on rules and principles incompatible with local circumstances replaced most of the African customary laws. The Received English Laws were prominent among the supplanting European laws, which displaced African customary laws in the British-colonized parts of Africa. This replacement led to conflicts of laws in most African jurisdictions post-colonization, raising questions about which of the dual systems of law should govern several transactions, especially among Africans and non-Africans.

Notwithstanding that European rules and principles dominate criminal procedures in most African courts, certain Africans still prefer to be ruled by their traditional law. This right to be governed by customary law faces further erosion in the face of AI tools trained with English law-informed statutes on criminal trials and sentencing. Without change, AI sentencing technology predicated on this status quo can potentially be discriminatory if it must compulsorily sentence a person under an imposed legal regime.

(5) Infrastructural Deficits

Closely related to the above-discussed points and an indirect impact of colonialism on Africa is the extant situation of infrastructural deficits, such as the limited investment in digitized case records, low-speed internet connectivity, scarce cybersecurity infrastructure, and unreliable access to power supply, all of which are critical for the successful implementation of AI systems. For example, while certain parts of the continent experience gross erratic power supply, some sub-Saharan countries suffer from total outages. According to a 2021 report, nearly forty-three percent of Africans have no access to power, with the majority in sub-Saharan Africa. While it transcends the precinct of this article, it is notable that centuries of economic deprivation coupled with bad governance systems that succeeded colonialism played key roles in current Africa’s power issues. Electric power issues impact residential buildings, private companies, and government parastatals, including courthouses. Most AI tools are bound to break down without electricity. They could consequently occasion injustice if they are constantly interrupted by power outages, presenting just one of the many infrastructural challenges to the adoption of AI in African courts. To make a point about the global resource divide, data centers alone consume about 1 percent of all global electricity, which is arguably more than the electricity consumed by half of Sub-Saharan Africa.

Quo Vadis?

Whilst algorithmic sentencing tools may propose some positive benefits to criminal justice systems, including how they could help achieve uniformity, reduce judicial biases and arbitrariness, increase efficiency and effectiveness, and promote better access to justice, peculiar and extant challenges exist for African courts. If it is inconceivable for some of these challenges to be resolved, they should, at the very least, be taken into account before AI tools are considered for deployment. Certain pre-colonial African values need to be restored, particularly the need to prioritize restorative over punitive justice. This requires a holistic amendment to criminal justice statutes and courts’ rules to accommodate the balancing of societal equilibrium as the main purpose of criminal sentencing. Perhaps this can be preferentially trained into the AI models that will be used to serve this purpose. Also, judicial corruption needs to be fought with a tough disposition. Perhaps AI tools could be developed instead to monitor court charges and improve reporting.

As much as court officers need to be trained on the utility of the AI tools deployed in their sector, those developing these AI systems also need to be trained on Africa’s judicial history. Even if deployed, suspicious court sentences should be subjected to higher and explainable review, and the right to contest bias from such AI systems should be recognized. Notwithstanding, many African judicial systems struggle with critical infrastructure such as electricity supply, telecommunications, computing, and even basic amenities. Several preliminary steps may need to be taken before African courts can experiment with the adoption of AI tools for criminal sentencing.