Agentic artificial intelligence (AI) represents a paradigm shift from current AI systems, as it can operate autonomously with minimal human oversight. Unlike generative AI that responds to prompts, agentic AI can independently set goals, create plans, and execute 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, “AI is beginning to help build better AI.” This self-accelerating loop could lead to capability development far outpacing current projections, warns Ylli Bajraktari, president of SCSP, in a recent newsletter.
In a global security context, Bajraktari emphasizes that “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, potentially using it for coercion, espionage, and influence operations.
Contrary to common policy assumptions, effective governance of agentic AI focuses not on the AI model itself but on the scaffolding built around it, SCSP experts explain. This scaffolding includes connectors to bridge the model to real-world infrastructure like email and financial platforms, memory for learning and adaptation, planning capabilities to break down objectives, permission structures defining system access, and guardrails that dictate refusals, such as spending limits or human sign-offs.
Accountability remains a critical challenge, with governance falling short in three key ways according to SCSP. First, responsibility is untraceable when AI agents act on behalf of users, making it impossible to determine who authorized what. Second, current frameworks only assess task completion, not whether the task was performed safely or caused harm. Third, agentic AI builds personal profiles that accumulate sensitive data on behavior patterns, preferences, and inferences, potentially exceeding what individuals wish to share.
Despite these challenges, SCSP experts stress that agentic AI is not to be feared or deferred. Institutions that prioritize understanding, shaping, and governing this technology will determine their competitive position and influence the global environment in which agentic AI operates. For more insights on how the United States can pursue effective governance, visit scsp.ai.
