Agentic AI and The Rise of Impact Futures Markets: A New Paradigm for Project Value Creation

Achim von Heynitz, University of Erfurt, Germany

ABSTRACT

Traditional Results-Based Management (RBM) frameworks in international development organizations remain inherently constrained by linear control cycles and retrospective learning; managerial incentives are heavily biased toward short-term output delivery and disbursement over long-term sustainable outcomes. To address these structural misalignments, this paper introduces a conceptual paradigm shift centered on a human-in-the-loop Agentic AI Orchestration Hub. We propose three interlinked, mutually reinforcing algorithmic innovations that transform the project life cycle: (1) Algorithmic Triangulation to internalize accountability by continuously synthesizing output delivery, risk exposure, and assumption validity; (2) an Impact Futures Market (IFM) that converts future probabilistic outcomes into present-day decision signals via Tradable Impact Assets (TIAs) under adversarial EvalAgent validation; and (3) a Retrospective Impact Market (RIM) backed by contingent, flexible post-implementation credit lines to capture and value emergent, unplanned results. This distributed agent-supporfted framework transitions project management from deterministic plan execution to continuous, adaptable outcome stewardship

Keywords

Agentic AI, Teleological Orchestration in Results-Based Management (RBM), Tradable Impact Assets (TIAs), Impact Futures Markets, Multilateral Development Banks.

Problems With AI Training: Sources and Consequences of Hallucination in Neural Network Models

Alan Chickinsky, Life Senior Member, IEEE

ABSTRACT

Contemporary AI language models are increasingly deployed in safety-critical contexts, yet they remain prone to generating dangerous or factually incorrect outputs. Documented examples include AI-generated recipes suggesting the use of chlorine-producing ingredient combinations [1]. Rather than acknowledging these as model errors, developers have characterized them as “hallucinations”—a term that obscures the underlying technical causes. This paper investigates the structural and methodological sources of hallucination in neural network models, examines three predominant training paradigms, and argues that incomplete training data, flawed train–test partitioning, and the inaccessibility of expert tacit knowledge are root causes of unreliable AI outputs. Recommendations for more rigorous training methodologies are proposed.

Keywords

Artificial Intelligence, Hallucinations, Neural Networks.