Admissibility, Reliability and the Confrontation Clause in AI-Generated Evidence: Lessons for the Nigerian Judiciary

Authors

  • Tarms Kagbala Federal Polytechnic Ekowe, Bayelsa State

Keywords:

Fair hearing rights, admissibility, AI-generated evidence, hearsay, machine declaration

Abstract

Artificial intelligence (AI) has become increasingly significant in twenty-first-century criminal justice systems. Law-enforcement agencies now deploy facial-recognition systems, predictive policing tools, automated forensic analysis and algorithmic risk assessments. Nigerian institutions are beginning to explore similar technologies, often without parallel development in evidentiary doctrine. This article examines the status of AI-generated evidence within Nigeria’s legal system and explores its implications for admissibility, reliability and the constitutional right to confront adverse evidence. Drawing upon developments in the United States, United Kingdom, European Union, South Africa and Kenya, the paper argues that Nigeria must adopt structured reliability standards, enforce disclosure obligations and strengthen judicial technological literacy. It concludes that without doctrinal clarity and institutional oversight, AI technologies risk undermining fair-hearing guarantees and eroding public confidence in the administration of justice.

References

Bullcoming v New Mexico 564 US 647 (2011).

Crawford v Washington 541 US 36 (2004).

Daubert v Merrell Dow Pharmaceuticals Inc 509 US 579 (1993).

Kubor v Dickson (2013) 4 NWLR (Pt 1345) 534.

Kubor v Dickson (2013) 4 NWLR (Pt 1345) 534.

Oshodin v State (2012) LPELR-7820 (CA).

R v T [2011] 1 Cr App R 9.

S v Mkohle 1990 (1) SACR 95 (A).

Tsalibawa v Habiba (1991) 2 NWLR (Pt 174) 461.

Legislation

Constitution of the Federal Republic of Nigeria 1999 (as amended).

Evidence Act 2011 (Nigeria).

European Convention on Human Rights (1950).

Books and Articles

Andrew Ashworth and Mike Redmayne, The Criminal Process (5th edn, OUP 2020).

Gary Edmond, ‘Forensic Science Evidence and the Limits of Cross-Examination’ (2015) 39 Crim LJ 198.

Joy Buolamwini and Timnit Gebru, ‘Gender Shades’ (2018) 81 Proceedings of Machine Learning Research 77.

Paul Roberts and Adrian Zuckerman, Criminal Evidence (3rd edn, OUP 2010).

Sandra Wachter, Brent Mittelstadt and Chris Russell, ‘Why Algorithms Are Not Neutral’ (2020) 34 Harvard JLT 1.

Reports

AI Now Institute, AI Now Report 2019.

National Research Council, Strengthening Forensic Science in the United States: A Path Forward (2009).

Downloads

Published

2026-02-11