Leveraging AI for the poor: a lens of problem-oriented innovation system

Purpose The multidimensional nature of poverty and its broad scope have caused the ineffectiveness of traditional approaches in dealing with poverty and the necessity of using innovative and digital solutions such as artificial intelligence (AI) technology. Based on the problem-oriented innovation system (PIS), this paper aims to focus on the role of AI in combat against poverty. Design/methodology/approach This paper has systematically reviewed the scientific articles focusing on the role of AI in the combat against poverty. After identifying 508 initial sources, 41 articles were selected according to their relevance to the research to address which functions of PIS have been neglected in previous studies and identify the literature gap that affects the efficiency of combating them. Findings The findings indicate that the literature mainly focuses on technology’s role, social entrepreneurship, regulation, promotion of public awareness and dissemination of results while paying less attention to standardization and provision of financial and human resources. Also, most of the studies have focused on the outcomes of AI application in poverty control and have not addressed the prerequisites of using AI for each PIS function. Focusing on the literature gap, the findings help direct future studies. Practical implications Identifying research gaps informs future research and prevents repetitive studies. Focusing on PISs and examining AI applications and functions, this paper has taken a step beyond existing knowledge and identified gaps in the literature regarding standardization procedures, budget allocation and human resources. Furthermore, presenting a structure based on prerequisites and outcomes of AI-aided poverty combat for each function has contributed to expanding the theoretical foundations of the literature. Social implications Reviewing articles focused on each function and their frequency has provided new ideas and guidelines for scholars to explore functions that have received less attention. The causal relationships between functions are also worth exploring, and future studies can focus on this aspect. Originality/value This paper focuses on a new perspective on AI and poverty, suggesting that AI assistance should not be limited to identifying and measuring poverty. Instead, the efficiency of AI applications maximizes when other aspects of the grand challenge (e.g. regulation, standardization, resource allocation, human resources and knowledge dissemination) are pursued, which have been neglected in previous studies on AI-aided poverty control.