Details

The purpose of this policy brief is to determine how governments can embed the human-in-the-loop (HITL) principle to advance more responsible AI. With the introduction of AI agents, exponential data growth, faster computing hardware, and complex, interconnected third-party developers, the ability for humans to manage or interfere with AI’s autonomous decisions is decreasing. The paper highlights key challenges of HITL from definition to implementation. The research focuses on the lifecycle of an AI system and human actors along the value chain. It identifies gaps that are opportunities for policymakers to step in for AI good governance. Further, the recommendations provide guidelines for governments, industry, society, nonprofits, and academia on how to work together to build, fund, endorse, adopt, deploy, use, and decommission responsible AI with adequate human agency. The policy refers to governments researching, working with, contracting with, endorsing, deploying, or vetting AI systems within their jurisdictions.