Legal and Economic Ecosystem Transformation Through AI: Innovation Challenges for Human-Centric Development

The current technological shift towards the application of artificial intelligence (AI) is gradually reshaping the way in which the legal and economic ecosystem functions, particularly in the agricultural sector. This study addresses agricultural economics, policies and rural development in the European Union (EU) and in the Republic of Serbia as a candidate for EU membership. It draws on a structured review of the academic literature and of EU, international and Serbian policy documents issued over the past ten years, including the AI Act, the Digital Omnibus on AI, the Data Governance Act and the proposal for the Common Agricultural Policy after 2027. The empirical component is an exploratory cross-sectional survey of 150 respondents drawn from three professional groups—farmers, lawyers and engineers. Four research hypotheses on the perceived human-centric development potential of AI applications were tested; one was confirmed, two were partially confirmed, and one was not confirmed. Major transformation channels across six regulatory dimensions of the legal framework were mapped onto agricultural AI use cases, showing that the human-centric safeguards of EU law apply with uneven effect: farmer credit scoring, seasonal labour management, the assessment of eligibility for public benefits and AI embedded in machinery fall under high-risk obligations, whereas agronomic decision support systems are governed mainly by data protection law, the Data Act and contractual arrangements. The survey results show that the perceived human-centric benefits of AI differ across the three groups, with farmers reporting the lowest expected benefits and the highest perceived risks of exclusion and labour displacement. Trust in institutions and legal protection was the only construct significantly associated with expected benefits once professional group and other perceptions were controlled, and the differences between the groups were largely accounted for by differences in trust, AI literacy and perceived data sovereignty. Several innovation challenges were identified, the principal ones being the generational and digital divide, labour transformation, platform concentration and institutional capacity; these are particularly acute in Serbia, where the vast majority of holdings are family farms. The proposed Human-Centric Agri-Rural AI Transformation (HCAT) framework links these challenges to four governance pillars, twelve instruments and twelve monitoring indicators, which are offered as a conceptual contribution to be validated in future research, and it provides a sequenced approach for candidate countries aligning with EU policies and frameworks.