The rise of agentic artificial intelligence—systems capable of autonomous goal-setting, planning, and execution with minimal human oversight—is reshaping the landscape of AI governance. Unlike traditional AI that responds to prompts, agentic AI can independently set objectives, create plans, and carry out multi-step tasks. According to experts at the Special Competitive Studies Project (SCSP), a nonprofit and nonpartisan initiative focused on strengthening America's long-term AI competitiveness, this evolution marks a significant shift: "AI is beginning to help build better AI."
Ylli Bajraktari, president of SCSP, warned in a recent newsletter that a self-accelerating loop—where AI capability improves and AI development compounds—could lead to capability growth far outpacing current projections. He emphasized the security implications: "An agent that can navigate complex bureaucratic systems, identify exploitable vulnerabilities, and act without leaving a clear attribution trail represents a qualitative expansion of adversarial capability." The United States must recognize that adversaries will deploy agentic AI in areas where governance is weakest, using it for coercion, espionage, and influence operations.
Effective governance, SCSP experts argue, does not center on the AI model itself but on the "scaffolding" built around it. This scaffolding includes connectors that bridge the model to real-world infrastructure such as email, booking systems, and financial platforms; memory that enables learning and adaptation across interactions; planning capabilities that break large objectives into smaller tasks and navigate obstacles; permission structures defining system access and actions; and guardrails determining what the system will refuse to do, such as spending limits or human sign-offs.
Accountability remains a critical challenge, with governance falling short in three key areas. First, responsibility is untraceable when AI acts autonomously—there is no clear way to determine who authorized what. Second, current frameworks only assess whether a task was completed, not whether it was performed safely or caused harm. Third, agentic AI builds personal profiles that may include sensitive data by accumulating information on behavior patterns, preferences, and inferences, often beyond what individuals intend to share.
Despite these challenges, SCSP emphasizes that agentic AI is not a technology to be feared. Institutions that prioritize understanding, shaping, and governing it will determine their competitive position and influence the global environment in which agentic AI operates. Proactive governance, focused on the scaffolding rather than the model, is essential to harness the benefits of agentic AI while mitigating risks.
