AI-driven portfolio optimisation with human behavioural adjustments for sustainable investment performance, Omotayo Bolarinde, Rasheedul Haque

Omotayo Bolarinde
DBA Research Scholar, Rushford Business School
MKH Properties and Supermarket
Nigeria
ORCID: 0009-0000-4793-3914
Dr Rasheedul Haque
School of Management and Business (SOMB)
MILA University, Nilai
Malaysia
ORCID: 0000-0001-8170-5413

Purpose: This research creates an AI-based portfolio optimisation model, which incorporates human behavioural knowledge and Artificial Intelligence (AI) applications to enhance sustainable investment outcomes. Although AI models like mean-variance optimisation and machine learning-based asset allocation are computationally efficient, they tend to overlook behavioural biases and real-world decision constraints of investors.
Design/Methodology: This study suggests a hybrid model in which AI would calculate ideal portfolio weights due to financial returns, volatility, Environmental, Social, and Governance (ESG) scores, and human intelligence would modify allocations based on risk perception, ethical concerns, and market sentiment. The process will include creating portfolios based on historical stock data and ESG ratings and optimisation with the Sharpe ratio and downside risk indicators. Portfolio performance measures will include portfolio return, standard deviation, beta, ESG-adjusted performance, and ESG-based performance. The human factors will be added by scenario-based adjustments to show the impact when the market becomes volatile or experiences shocks related to sustainability in a simulated manner. The research will compare the results of AI-only portfolios with those of hybrid portfolios in terms of risk-adjusted returns and sustainability alignment.
Findings/Results: The results indicate that, through a blend of machine accuracy and human judgment, there is a stronger and more moralised approach to investment.
Originality/Value: The originality of this study lies in combining AI-based optimisation with human behavioural insights and ESG-driven adjustments within a single framework. It extends conventional portfolio theory by embedding ethical considerations and sustainability risk responses into the optimisation process, which are typically absent in purely algorithmic models.
Practical Implications: The study provides a feasible solution to asset managers and institutional investors, as well as individual investors willing to incorporate sustainability in their portfolio decisions without compromising their financial performance.
Keywords: ESG; Artificial Intelligence; Sustainability; Finance.