AI and Corruption: Legal Liability in Algorithmic Decision-Making
Summary: 1. Introduction. – 2. Methodology. – 3. Results. – 3.1. AI Applications in Anti-Corruption. – 3.2. How AI Systems Perpetuate Bias and Discrimination. – 3.3. Current State of Liability. – 3.4. Responsibility Matrix. – 4. Discussions. – 5. Conclusions.
Background: The question of whether machines can be corrupt appears paradoxical; nevertheless, it is rapidly gaining relevance in the world of artificial intelligence (AI) and changing how decisions are made in public and government systems. These systems offer notable advantages, including enhanced efficiency, reduced human error, and the ability to combat corruption by detecting fraud, tracking funds, and improving public services. It can make decisions based on data instead of personal interests. However, the use of AI is not without risks. When trained on biased datasets, AI systems may produce unfair outcomes. Additionally, if AI systems are deliberately manipulated for personal or political gain, they may support or conceal corrupt actions. This research examines the role of AI in public services, exploring its potential to prevent or contribute to corruption. The goal is to understand where AI is safe and where it is risky.
Methods: The research used a qualitative research design. Data was collected by reviewing academic papers, laws, and official reports. Sources were identified using academic databases such as Google Scholar, with a focus on peer-reviewed law journals, policy briefs, and official government documents. All materials were checked using the CRAAP test. The method for analysing the data was doctrinal legal analysis.
Results and Conclusions: The findings indicate that AI has considerable potential to enhance transparency and reduce bribery by limiting human control in administrative processes. However, in countries with weak legal systems, AI can be misused. When AI systems lack transparency or explainability, they can obscure corrupt practices rather than expose them. This risk is pronounced in high-stakes domains such as public procurement and budgeting systems.
While certain countries have implemented robust legal safeguards and effective audits that mitigate risks, many others lack clear rules on who is responsible when AI contributes to corruption. In numerous cases, public AI systems lack external checks, and existing mechanisms for reporting corruption are not equipped to address AI-specific issues. As a result, accountability gaps persist.
The study highlights the continued importance of human oversight to stop manipulation. It recommends that governments strengthen regulatory frameworks by introducing explicity provisions on accountability. Independent audits should be added to all public AI systems. Whistleblower systems should be updated to accommodate AI-related cases.
Abstract
Background: The question of whether machines can be corrupt appears paradoxical; nevertheless, it is rapidly gaining relevance in the world of artificial intelligence (AI) and changing how decisions are made in public and government systems. These systems offer notable advantages, including enhanced efficiency, reduced human error, and the ability to combat corruption by detecting fraud, tracking funds, and improving public services. It can make decisions based on data instead of personal interests. However, the use of AI is not without risks. When trained on biased datasets, AI systems may produce unfair outcomes. Additionally, if AI systems are deliberately manipulated for personal or political gain, they may support or conceal corrupt actions. This research examines the role of AI in public services, exploring its potential to prevent or contribute to corruption. The goal is to understand where AI is safe and where it is risky.
Methods: The research used a qualitative research design. Data was collected by reviewing academic papers, laws, and official reports. Sources were identified using academic databases such as Google Scholar, with a focus on peer-reviewed law journals, policy briefs, and official government documents. All materials were checked using the CRAAP test. The method for analysing the data was doctrinal legal analysis.
Results and Conclusions: The findings indicate that AI has considerable potential to enhance transparency and reduce bribery by limiting human control in administrative processes. However, in countries with weak legal systems, AI can be misused. When AI systems lack transparency or explainability, they can obscure corrupt practices rather than expose them. This risk is pronounced in high-stakes domains such as public procurement and budgeting systems.
While certain countries have implemented robust legal safeguards and effective audits that mitigate risks, many others lack clear rules on who is responsible when AI contributes to corruption. In numerous cases, public AI systems lack external checks, and existing mechanisms for reporting corruption are not equipped to address AI-specific issues. As a result, accountability gaps persist.
The study highlights the continued importance of human oversight to stop manipulation. It recommends that governments strengthen regulatory frameworks by introducing explicity provisions on accountability. Independent audits should be added to all public AI systems. Whistleblower systems should be updated to accommodate AI-related cases.
About Authors
Authors information
Naeem AllahRakha
Ph.D. (Law), Faculty of Law, Department of Cyber Law, Tashkent State University of Law, Tashkent, Uzbekistan
https://orcid.org/0000-0003-3001-1571
Corresponding author, solely responsible for conceptualization, data curation, formal analysis, funding acquisition, methodology, resources, validation and writing – original draft.
Competing interests: No competing interests were disclosed.
Disclaimer: The author declares that his opinion and views expressed in this manuscript are free of any impact of any organizations.
Rights and Permissions
Copyright: © 2025 Naeem AllahRakha. This is an open access article distributed under the terms of the Creative Commons Attribution License, (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Editors
Managing editor – Mag. Yuliia Hartman. English Editor – Julie Bold. Ukrainian Language Editor – Liliia Hartman.
References
1. Alhosani K and Alhashmi SM, ‘Opportunities, Challenges, and Benefits of AI Innovation in Government Services: A Review’ (2024) 4(1) Discover Artificial Intelligence 18, doi:10.1007/s44163-024-00111-w.
2. AllahRakha N, ‘Global Perspectives on Cybercrime Legislation’ (2024) 8(10) Journal of Infrastructure, Policy and Development 6007, doi:10.24294/jipd.v8i10.6007.
3. AllahRakha N, ‘The Legality of Reverse Engineering and the Protection of Trade Secrets in the Software Industry’ (2025) 15(2) Jurisdictie: Jurnal Hukum dan Syariah 309, doi:10.18860/j.v15i2.28422.
4. Alshahrani A and others, ‘Artificial Intelligence and Decision-Making in Government Functions: Opportunities, Challenges and Future Research’ (2024) 18(4) Transforming Government: People, Process and Policy 678, doi:10.1108/TG-06-2024-0131.
5. Brahmia N, ‘The Role of Artificial Intelligence in Enhancing Tax Compliance and Customs Efficiency: A Case Study of the South African Revenue Service (SARS)’ (2025) 19(5) International Journal of Economic Perspectives 1881.
6. Brożek B and others, ‘The Black Box Problem Revisited: Real and Imaginary Challenges for Automated Legal Decision Making’ (2024) 32(2) Artificial Intelligence and Law 427, doi:10.1007/s10506-023-09356-9.
7. Cheong BC, ‘Transparency and Accountability in AI Systems: Safeguarding Wellbeing in the Age of Algorithmic Decision-Making’ (2024) 6 Frontiers in Human Dynamics 1421273, doi:10.3389/fhumd.2024.1421273.
8. Cho TN, Look V and Silver HJ, ‘Recognizing the Primacy of Artificial Intelligence in America: Biden’s Executive Order Sets a High Bar for Regulation and Innovation’ (Womble Bond Dickinson, 10 November2023) <https://www.womblebonddickinson.com/us/insights/alerts/recognizing-primacy-artificial-intelligence-america-bidens-executive-order-sets> accessed 5 July 2025.
9. Cousineau C, Dara R and Chowdhury A, ‘Trustworthy AI: AI Developers’ Lens to Implementation Challenges and Opportunities’ (2025) 9(2) Data and Information Management 100082, doi:10.1016/j.dim.2024.100082.
10. Dastani M and Yazdanpanah V, ‘Responsibility of AI Systems’ (2023) 38 AI & SOCIETY 843, doi:10.1007/s00146-022-01481-4.
11. De Carvalho ÁAB and Filho RH, ‘Detecting Fraud in Public Acquisition of Brazilian Government with an Analytical Approach’ (2024) 238 Procedia Computer Science 248, doi:10.1016/j.procs.2024.06.022.
12. De Silva D and Alahakoon D, ‘An Artificial Intelligence Life Cycle: From Conception to Production’ (2022) 3(6) Patterns 100489, doi:10.1016/j.patter.2022.100489.
13. Enqvist L, ‘“Human Oversight” in the EU Artificial Intelligence Act: What, When and by Whom?’ (2023) 15(2) Law, Innovation and Technology 508, doi:10.1080/17579961.2023.2245683.
14. Ferrara E, ‘Fairness and Bias in Artificial Intelligence: A Brief Survey of Sources, Impacts, and Mitigation Strategies’ (2023) 6(1) Sci 3, doi:10.3390/sci6010003.
15. Garcia-Segura LA, ‘The Role of Artificial Intelligence in Preventing Corporate Crime’ (2024) 5 Journal of Economic Criminology 100091, doi:10.1016/j.jeconc.2024.100091.
16. Gillespie T, ‘Content Moderation, AI, and the Question of Scale’ (2020) 7(2) Big Data & Society 2053951720943234, doi:10.1177/2053951720943234.
17. Guida M and others, ‘The Role of Artificial Intelligence in the Procurement Process: State of the Art and Research Agenda’ (2023) 29(3) Journal of Purchasing and Supply Management 100823, doi:10.1016/j.pursup.2023.100823.
18. Judge B, Nitzberg M and Russell S, ‘When Code Isn’t Law: Rethinking Regulation for Artificial Intelligence’ (2025) 44(1) Policy and Society 85, doi:10.1093/polsoc/puae020.
19. Kattel R and Mergel I, ‘Estonia’s Digital Transformation: Mission Mystique and the Hiding Hand’ in 't Hart P and Compton M (eds), Great Policy Successes (OUP 2019) 143, doi:10.1093/oso/9780198843719.003.0008.
20. Kim TW and Routledge BR, ‘Why a Right to an Explanation of Algorithmic Decision-Making Should Exist: A Trust-Based Approach’ (2022) 32(1) Business Ethics Quarterly 75, doi:10.1017/beq.2021.3.
21. Kostka G, ‘China’s Social Credit Systems and Public Opinion: Explaining High Levels of Approval’ (2019) 21(7) New Media & Society 1565, doi:10.1177/14614448198264.
22. Laato S and others, ‘How to Explain AI Systems to End Users: A Systematic Literature Review and Research Agenda’ (2022) 32(7) Internet Research 1, doi:10.1108/INTR-08-2021-0600.
23. Leib M and others, ‘Corrupted by Algorithms? How AI-Generated and Human-Written Advice Shape (Dis)Honesty’ (2024) 134(658) The Economic Journal 766, doi:10.1093/ej/uead056.
24. Lima MSM and Delen D, ‘Predicting and Explaining Corruption across Countries: A Machine Learning Approach’ (2020) 37(1) Government Information Quarterly 101407, doi:10.1016/j.giq.2019.101407.
25. Lopez Acera A, ‘Artificial Intelligence and the Fight against Corruption’ (Agency for the Prevention and Fight against Fraud and Corruption of the Valencian Community, 21 November 2023) <https://www.antifraucv.es/en/artificial-intelligence-and-the-fight-against-corruption/> accessed 5 July 2025.
26. Menke W, Gomes R and Xavier F, ‘Impacts of AI-Based Anti-Corruption Audits on Risk Aversion in Decision-Making: A Case Study of the Brazilian ALICE Tool’ (2024) 4 Global Public Policy and Governance 273, doi:10.1007/s43508-024-00098-1.
27. Montagnani ML, Najjar MC and Davola A, ‘The EU Regulatory Approach(Es) to AI Liability, and Its Application to the Financial Services Market’ (2024) 53 Computer Law & Security Review 105984. doi:10.1016/j.clsr.2024.105984.
28. Nurse A, ‘Law Enforcement’, Oxford Research Encyclopedia of Criminology (OUP 2024) doi:10.1093/acrefore/9780190264079.013.760.
29. Odilla F, ‘Bots against Corruption: Exploring the Benefits and Limitations of AI-Based Anti-Corruption Technology’ (2023) 80 Crime, Law and Social Change 353, doi:10.1007/s10611-023-10091-0.
30. Sadok H, Sakka F and El Maknouzi MEH, ‘Artificial Intelligence and Bank Credit Analysis: A Review’ (2022) 10(1) Cogent Economics & Finance 2023262, doi:10.1080/23322039.2021.2023262.
31. Sanger R, ‘Artificial Intelligence and Criminal Law’ (The Colleges of Law, 12 January 2024) <https://www.collegesoflaw.edu/blog/2024/01/12/artificial-intelligence-and-criminal-law/> accessed 5 July 2025.
32. Tveita LJ and Hustad E, ‘Benefits and Challenges of Artificial Intelligence in Public Sector: A Literature Review’ (2025) 256 Procedia Computer Science 222, doi:10.1016/j.procs.2025.02.115.
33. Vuković DB, Dekpo-Adza S and Matović S, ‘AI Integration in Financial Services: A Systematic Review of Trends and Regulatory Challenges’ (2025) 12(1) Humanities and Social Sciences Communications 562, doi:10.1057/s41599-025-04850-8.
34. Wang Z and others, ‘A Red Team Automated Testing Modeling and Online Planning Method for Post-Penetration’ (2024) 144(11) Computers & Security 103945, doi:10.1016/j.cose.2024.103945.
35. Wood JM and Spagnolo A, ‘Implementing Effective AI Compliance amid an Evolving Regulatory Landscape’ (GIR, 7 August 2024) <https://globalinvestigationsreview.com/guide/the-guide-compliance/third-edition/article/implementing-effective-ai-compliance-amid-evolving-regulatory-landscape> accessed 5 July 2025.
36. Zacarés JM, ‘The Iron Triangle of Urban Entrepreneurialism: The Political Economy of Urban Corruption in Spain’ (2020) 52(5) Antipode 1351, doi:10.1111/anti.12637.
Reviews for article
Add a Review
АНОТАЦІЯ УКРАЇНСЬКОЮ МОВОЮ
Дослідницька стаття
ШІ ТА КОРУПЦІЯ: ЮРИДИЧНА ВІДПОВІДАЛЬНІСТЬ В АЛГОРИТМІЧНОМУ ПРИЙНЯТТІ РІШЕНЬ
Наїм АллагРаха
АНОТАЦІЯ
Вступ. Питання про те, чи можуть машини бути корумпованими, здається парадоксальним; проте воно швидко набуває актуальності у світі штучного інтелекту (ШІ) та змінює те, як приймаються рішення в державних та урядових системах. Ці системи пропонують помітні переваги, зокрема підвищену ефективність, зменшення людських помилок та здатність боротися з корупцією за допомогою виявлення шахрайства, відстеження коштів та покращення державних послуг. Він може приймати рішення на основі даних, а не особистих інтересів. Однак використання ШІ не позбавлене ризиків. Під час навчання на упереджених наборах даних системи ШІ можуть призводити до несправедливих результатів. Крім того, якщо системами ШІ навмисно маніпулювати для особистої чи політичної вигоди, вони можуть підтримувати або приховувати корупційні дії. У цій статті розглядається роль ШІ в державній службі, вивчається його потенціал у запобіганні або сприянні корупції. Мета полягає в тому, щоб зрозуміти, де ШІ безпечний, а де ризикований.
Методи. У дослідженні використовувався якісний дизайн дослідження. Дані були зібрані за допомогою огляду академічних робіт, законів та офіційних звітів. Джерела були визначені за допомогою академічних баз даних, таких як Google Scholar, з наголосом на рецензованих юридичних журналах, аналітичних звітах та офіційних урядових документах. Усі матеріали були перевірені за допомогою тесту CRAAP. Аналіз правової доктрини також використовувався як метод.
Результати та висновки. Результати вказують на те, що ШІ має значний потенціал для підвищення прозорості та зменшення хабарництва, якщо обмежити людський контроль в адміністративних процесах. Однак у країнах зі слабкими правовими системами може бути зловживання штучним інтелектом. Коли системам ШІ бракує прозорості або пояснень, вони можуть приховувати корупційні практики, а не викривати їх. Цей ризик яскраво виражений у сферах з високими ставками, таких як системи державних закупівель та бюджетування. Хоча деякі країни запровадили надійні правові гарантії та ефективні аудити, які знижують ризики, багатьом іншим бракує чітких правил щодо того, хто несе відповідальність, коли ШІ сприяє корупції. У багатьох випадках державні системи ШІ не мають зовнішніх перевірок, а наявні механізми повідомлення про корупцію не пристосовані для вирішення проблем, пов'язаних зі ШІ. Як наслідок, прогалини у підзвітності досі є. У дослідженні підкреслено постійну важливість людського нагляду для припинення маніпуляцій. Також було рекомендовано урядам зміцнити нормативно-правову базу, ввівши чіткі положення про підзвітність. Незалежні аудити слід додати до всіх публічних систем штучного інтелекту. Системи інформування про порушення слід оновити, щоб враховувати випадки, пов'язані зі штучним інтелектом.
Ключові слова: штучний інтелект (ШІ), корупція, юридична відповідальність, алгоритм, прийняття рішень за допомогою ШІ.
Publication history
-
DETAILS FOR PUBLICATION
Date of submission: 18 Jun 2025
Date of acceptance: 23 Jul 2025
Last Publication: 17 Aug 2025
Whether the manuscript was fast tracked? - No
Number of reviewer reports submitted in the first round: 2 reports
Number of revision rounds: 1 round
Technical tools were used in the editorial process:
Plagiarism checks - Turnitin from iThenticate https://www.turnitin.com/products/ithenticate/
Scholastica for Peer Review https://scholasticahq.com/law-reviews
How to cite it?
-
AllahRakha N, ‘AI and Corruption: Legal Liability in Algorithmic Decision-Making’ (2025) 8(3) Access to Justice in Eastern Europe 303-264 <https://doi.org/10.33327/AJEE-18-8.3-a000120>
Managing editor – Mag. Yuliia Hartman. English Editor – Julie Bold. Ukrainian Language Editor – Liliia Hartman.
