The emergence of agentic artificial intelligence—systems capable of operating autonomously with minimal human oversight—is reshaping the landscape of global security and technological competition. Unlike conventional AI that responds to prompts, agentic AI can independently set goals, create plans, and execute multi-step tasks, potentially creating a self-accelerating loop where AI helps build better AI. According to experts at the Special Competitive Studies Project (SCSP), a nonprofit initiative focused on strengthening America's long-term AI competitiveness, this rapid capability development will far outrun current projections.
Ylli Bajraktari, president of SCSP, warned in a recent newsletter that agentic AI systems could be deployed by adversaries in areas where governance is weakest, enabling coercion, espionage, and influence operations. "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," he emphasized. This underscores the urgency for the United States to proactively shape the governance of agentic AI.
Effective governance, however, does not center on the AI model itself but on the scaffolding built around it, SCSP experts explain. This scaffolding includes connectors that bridge the model to real-world infrastructure such as email and financial platforms, memory that allows the system to learn and adapt over time, planning capabilities that break large objectives into smaller tasks, permission structures that define system access, and guardrails that determine what the system will refuse to do. Without robust scaffolding, agentic AI poses significant risks.
Accountability remains a critical challenge. Current frameworks fall short in three key ways: responsibility is untraceable when AI acts on behalf of a user; existing assessments focus only on task completion, not on whether the AI performed safely or caused harm; and agentic AI builds personal profiles that accumulate sensitive data on behavior and preferences. These gaps leave individuals and institutions vulnerable to unintended consequences.
Despite these challenges, SCSP emphasizes that agentic AI is not a technology to be feared or deferred. Institutions that prioritize understanding, shaping, and governing agentic AI will determine their competitive position and influence the global environment in which these systems operate. To learn more about how the United States should pursue effective governance, visit scsp.ai.
