Survey evidence from Nigeria suggests AI adoption in agriculture and waste-to-energy is linked to noticeable gains in operational efficiency and environmental sustainability. But infrastructure shortfalls, unstable power, limited technical expertise and high costs remain major obstacles to wider impact.

Main Finding

AI adoption in Nigeria’s agriculture and waste-to-energy (WTE) systems is moderately high and delivers clear productivity and sustainability benefits: survey evidence (n ≈ 522) shows AI adoption (mean = 3.84/5) is associated with higher operational efficiency (mean = 4.02) and improved environmental sustainability (mean = 3.95). Econometric results find AI investment significantly predicts operational efficiency (β = 0.62, p < 0.01) and sustainability outcomes (β = 0.55, p < 0.01). Respondents view long‑term benefits as outweighing implementation costs, despite barriers (infrastructure, electricity, skills, upfront cost).

Key Points

• Sample & reliability
  • Quantitative survey across Nigeria’s six geopolitical zones (states: Lagos, Rivers, Kano, Enugu, Plateau, Borno); reported sample ≈ 522 respondents; instrument Cronbach’s α = 0.87.
• Descriptive outcomes (5‑point Likert means)
  • AI adoption: 3.84
  • Operational efficiency: 4.02
  • Environmental sustainability: 3.95
  • Social impact: 3.78
  • Cost of implementation: 3.67
  • Technical capacity (skills): 3.41 (lowest)
• Regression results
  • AI investment → operational efficiency: β = 0.62 (SE = 0.08), t = 7.75, p < 0.001
  • AI investment → sustainability index: β = 0.55 (SE = 0.09), t = 6.11, p < 0.001
  • Other reported covariate: operational efficiency coefficient 0.48 (SE = 0.07, p < 0.001) in models presented
• Main barriers identified
  • Electricity supply instability, limited technical expertise, high initial costs, inadequate digital/internet infrastructure, policy gaps.
• Determinants of effective deployment (means)
  • Electricity stability (4.22), funding availability (4.18), technical skills (4.12), internet/digital infrastructure (4.05), government policy support (3.96), AI training (3.88), PPPs (3.91).
• Theoretical framing
  • Technology Acceptance Model, Cost–Benefit Theory (sustainability‑adjusted), Sustainable Development Theory.

Data & Methods

• Design: Cross-sectional quantitative survey plus econometric analysis (descriptive statistics and multiple regression).
• Coverage: Six states representing Nigeria’s geopolitical zones; respondents included farmers, waste managers, agricultural experts, tech/policy personnel.
• Sample procedure: Stratified random sampling targeting 600; ~87% response rate (reported final n ≈ 522).
• Instrument: Structured questionnaire with five sections on AI adoption, technological readiness, operational efficiency, energy & environmental sustainability, implementation challenges; 5‑point Likert scale; pilot (n = 50).
• Validity/robustness: Content validity via expert review; diagnostics reported (multicollinearity, normality, heteroscedasticity) prior to regression estimation. Regression tables report coefficients, SEs, t-values, p-values; R² and formal endogeneity checks not reported in the paper.
• Limitations in methods (implied/noted): Self‑reported outcome measures, cross‑sectional design limits causal inference; no engineering-level techno-economic model or lifecycle energy accounting presented.

Implications for AI Economics

• Returns and productivity
  • Survey and regression evidence suggest strong positive associations between AI investment and operational efficiency—supporting claims that AI can raise agricultural productivity and improve WTE conversion efficiency in low‑income settings.
• Value of incorporating externalities
  • Authors argue benefits include reduced waste, higher energy recovery, and environmental gains; researchers/economists should quantify these externalities (GHG reductions, avoided remediation costs) in monetary terms for full welfare comparisons.
• Policy and public finance
  • High upfront costs and infrastructure constraints imply a role for public policy: targeted subsidies, concessional finance, public investment in electricity/digital infrastructure, and support for training/extension to increase adoption among smallholders.
• Design of incentives and market interventions
  • PPPs, blended finance, pay‑for‑performance contracts (e.g., energy‑recovery revenue sharing), and service‑based delivery (AI-as-a-service, cooperative access) can lower entry costs and improve diffusion.
• Research priorities for stronger economic evidence
  • Move from cross‑sectional perception surveys to:
    • Controlled pilot trials or randomized evaluations measuring physical outputs (yields, energy recovered) and cost data.
    • Techno‑economic and lifecycle assessments to capture embodied and operational energy (including AI compute costs) and net GHG impacts.
    • Distributional analyses to assess effects on smallholders, labor markets, and welfare.
    • Dynamic models of scaling (cost declines, learning-by-doing) and sensitivity to electricity reliability and data/infrastructure constraints.
• Cautionary notes
  • Potential reverse causality/endogeneity: better firms/operations may both invest more in AI and be more efficient, so causal magnitudes may be overstated without stronger identification.
  • Computational energy use: AI systems themselves consume electricity — net sustainability gains depend on system design, energy source mix, and efficiency of models deployed.
• Bottom line for AI economics
  • This paper provides indicative evidence that AI investments in agriculture and WTE can generate measurable productivity and sustainability gains in a developing country context. For policy and investment decisions, more granular, impact‑oriented economic evaluations (including lifecycle energy and distributional effects) are needed to move from suggestive associations to robust cost‑benefit and policy prescriptions.