Ortopah Kojo Botchey

Dr. Ortopah Kojo Botchey

orcid

Research Focus & Metrics

Academic Career Length

1Year (2026-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
19 April 2026 in Businesses
Verified
Journal Article
Open access
PDF file
Victor FrimpongVictor Frimpong
Victor Frimpong
OBOrtopah Kojo Botchey
Ortopah Kojo Botchey
Purpose: This research investigates where formal responsibility for artificial intelligence (AI) lies within organisations and how the presence, absence, or structure of that responsibility affects their ability to govern AI effectively. Method: The study surveys 351 organisations across sectors and regions to examine AI governance roles. It focuses on authority, resources, and organisational integration, using hierarchical cluster analysis to identify governance configurations. Findings: The results indicate that formal AI governance roles are unevenly distributed and often weakly integrated into organisational structures. When these roles exist, they are usually placed below the executive level, lack sufficient authority, and differ greatly in the resources available to them. A cluster analysis reveals four governance configurations—Governance Absence, Symbolic Governance, Operational Governance, and Institutionalised Governance, indicating that governance capacity is primarily influenced by how well these roles are embedded in the structure, rather than just their presence. Implications: The findings suggest that AI governance may be better understood as a structural and organisational design issue, with potential implications for accountability and oversight. However, the relationship between governance configurations and outcomes, such as ethical risk and compliance, remains an area for future research. Originality: The study takes an absence-based approach to AI ethics, establishing a baseline for future research on governance maturity, compliance, trust, and ethical risk.

23 February 2026 in Preprints.org
Verified
Preprint
Open access
PDF file
Victor FrimpongVictor Frimpong
Victor Frimpong
OBOrtopah Kojo Botchey
Ortopah Kojo Botchey
As AI technologies increasingly play a crucial role in organisational decision-making, ethical frameworks and governance guidelines have been developed to ensure accountability, transparency, and responsible use. However, these governance structures primarily assume that organisations have the formal capacity to oversee AI, without examining whether such capacity is actually present. Empirical evidence on how organisations truly govern AI—and where responsibility is fundamentally lacking—remains scarce. This paper offers an initial empirical delineation of formal AI governance responsibilities across diverse sectors and regions. It employs survey data from 351 organisations to investigate the existence of positions such as Chief Artificial Intelligence Officer (CAIO), AI Ethics Officer, Responsible AI Lead, Algorithmic Auditor, and AI Governance Committees. Furthermore, it analyses variations in these jobs across industries and geographies, as well as their structural characteristics, such as seniority, reporting relationships, authority, and available resources. The research reveals prevalent profiles of governance maturity. The findings indicate that formal roles for AI governance are not consistently implemented and, when they do exist, often lack the necessary authority, resources, and integration at a senior institutional level. Executive-level leadership roles and specialised audit functions are rare, and many organisations operate without any formal AI governance roles despite using AI technologies. The study outlines four profiles of governance maturity: Governance Absence, Symbolic Governance, Operational Governance, and Institutionalised Governance, highlighting that mature governance is often more the exception than the norm. By empirically assessing the presence or absence of AI governance, this research presents an absence-based viewpoint on AI ethics. It indicates that ethical concerns often arise from inadequacies in governance design rather than from flaws in existing frameworks. These results establish a foundational empirical baseline for subsequent studies on how various AI governance models influence compliance, trust, and ethical risks.

17 March 2026 in Preprints.org
Verified
Preprint
Open access
PDF file
OBOrtopah Kojo Botchey
Ortopah Kojo Botchey
Technology adoption theories developed in institutionally mature contexts assume stable hierarchies among determinants, with perceived usefulness typically dominating. This paper qualifies this assumption by proposing that adoption hierarchies are institutionally contingent. Drawing on institutional voids theory and digital finance research, the paper develops a framework identifying three adoption regimes that function as ideal types which may overlap within contexts: (a) institutional trust dominant, where strong market supporting institutions enable usefulness-centered adoption; (b) vendor trust compensatory, where institutional voids elevate vendor-specific trust to primary importance; and (c) infrastructure-constrained, where basic access functions as a direct behavioral determinant. The framework extends technology acceptance theory by specifying when hierarchies change, theorizing trust as a compensatory mechanism, infrastructure as a hard constraint based on physical feasibility rather than perceptions, and a digital leveling effect explaining selective cultural influence. We derive propositions and outline a research agenda for cross-country and longitudinal validation, with implications for technology acceptance theory and digital financial inclusion practice.

29 July 2026 in SSRN Electronic Journal
Verified
Preprint
Open access
OBOrtopah Kojo Botchey
Ortopah Kojo Botchey
Ghana's National Artificial Intelligence Strategy 2025-2035, released by the Ministry of Communication, Digital Technology and Innovations on 12 December 2025 and publicly launched in Accra on 24 April 2026, commits the country to building AI that understands local languages, reflects local culture and is trained on Ghanaian data (Government of Ghana, 2025). Pillar 6 of the Strategy targets sector adoption, identifying insurance among its priority domains. This paper translates that ambition into an operational blueprint for the insurance sector. The argument is empirical, regulatory, theoretical and practical. Empirically, doctoral research on digital insurance adoption in Ghana finds that, when tested simultaneously, consumer trust is a stronger predictor of adoption attitudes than perceived usefulness, that ease of use is non-significant, and that infrastructure exerts a direct effect on adoption independent of attitudes and intentions (Botchey, 2025a). Theoretically, the paper extends the vendor trust compensatory regime (Botchey, 2025b) to AI-mediated insurance interactions, drawing on the established literatures on institutional voids (Khanna & Palepu, 1997, 2010), trust-integrated technology acceptance (Gefen, Karahanna, & Straub, 2003; Pavlou, 2003) and the unified theory of acceptance and use of technology (Venkatesh, Morris, Davis, & Davis, 2003), and reframes AI adoption in Ghanaian insurance as a trustengineering problem rather than a capabilities problem. Regulatorily, the existing supervisory framework under the Insurance Act, 2021 (Act 1061) and the National Insurance Commission is shown to be a sufficient platform for AI integration, provided that sector-specific guidance is issued in parallel with the forthcoming Emerging Technologies Bill and informed by international precedents including the European Insurance and Occupational Pensions Authority's opinion on AI governance (EIOPA, 2025) and the Financial Stability Institute's comparative review of AI in finance (FSI, 2024). Practically, the paper proposes a five-layer blueprint covering language and data, products, governance, skills and measurement. It also identifies a structural gap between the Strategy's ambition and the demonstrated readiness of public-facing AI systems, drawing on the publicly discussed launch of the Aku assistant to argue that local private-sector actors have a comparative advantage in implementation. The contribution is a regulator-implementable roadmap that aligns the Strategy's national objectives with the operational realities of an industry whose gross-premium penetration was 1.0 per cent of GDP in 2024 and 0.63 per cent on an IFRS 17 Insurance Service Revenue basis, highlighting that AI-enabled distribution must address structural under-penetration rather than only efficiency at current scale. Although the empirical setting is Ghana, the blueprint and the underlying regime classification are intended to be portable: the analytical pattern of Regime B and Regime C is encountered in many emerging-market insurance jurisdictions, and the paper aims to offer a transferable template for regulators in comparable institutional settings. The paper concludes with policy recommendations addressed to the National Insurance Commission, the Responsible AI Authority, insurers and brokers, and the Actuarial Society of Ghana.

1 January 2026 in SSRN Electronic Journal
Verified
Preprint
Open access
OBOrtopah Kojo Botchey
Ortopah Kojo Botchey
This study investigates how institutional voids reorder established technology adoption hierarchies in Ghana's digital insurance market. Drawing on a sequential mixed-methods design that integrates structural equation modelling of 226 digitally connected consumers with an organisational case study, the analysis tests whether trust and infrastructural conditions displace conventional Technology Acceptance Model (TAM) drivers. Trust in digital insurance platforms (β = 0.389, p < .001) emerges as the dominant antecedent of attitudes, surpassing perceived usefulness (β = 0.324, p < .001) and rendering perceived ease of use statistically non-significant once considered simultaneously (β = 0.096, p > .05). Technological infrastructure exerts a direct influence on adoption behaviour (β = 0.227, p < .001), bypassing the attitude-intention sequence typically foregrounded in TAM and UTAUT applications. Cultural dimensions show a differentiated pattern: masculinity retains a negative association with adoption attitudes, whereas power distance and uncertainty avoidance do not display significant effects in the integrated model. Collectively, the findings indicate that technology adoption hierarchies are contingent on institutional context: in the presence of institutional voids, vendor-specific trust can become primary, infrastructure operates as a binding constraint, and digital channels may attenuate some hierarchical cultural influences. The study contributes to digital transformation theory by demonstrating that adoption hierarchies are institutionally contingent rather than universal, with implications for financial inclusion strategies across emerging markets.

12 September 2026 in SSRN Electronic Journal
Verified
Preprint
Open access
OBOrtopah Kojo Botchey
Ortopah Kojo Botchey
On 27 July 2026, Regulation (EU) 2026/1744, the Digital Omnibus on Artificial Intelligence, entered into force, amending the EU Artificial Intelligence Act days before its general application date. The Regulation deferred the core Chapter III requirements for standalone high-risk AI systems under Article 6(2) and Annex III from 2 August 2026 to 2 December 2027, a revised timetable intended to align obligations with the availability of implementation support, including harmonised standards, guidance, and conformity assessment arrangements. This paper offers an early insurance-specific analysis of the deferral and makes three contributions. First, it maps the regulatory position of insurers as at August 2026, identifying an asymmetric application phase in which transparency obligations and the Act's governance and enforcement architecture apply while the substantive Annex III high-risk regime remains deferred. Second, it examines the scope asymmetry of Annex III, which expressly identifies AI intended for risk assessment and pricing in relation to natural persons in life and health insurance, while non-life underwriting, claims triage, fraud detection, and reserving are not expressly enumerated, despite carrying much of the weight of actuarial AI deployment. Third, it asks what emerging-market insurance regulators should learn from a benchmark jurisdiction that deferred its own timetable on the eve of application. Drawing on comparative supervisory approaches in Switzerland, the United States, and Ghana, and on the author's Trust-First framework of institutionally contingent technology adoption, the paper proposes a standards-readiness sequencing model gating application dates on norm availability, supervisory capacity, and market infrastructure. The deferral, the paper argues, is not free time but a governance window.