The 'use AI for safety' plan adopted by frontier labs is most likely to fail not because alignment techniques are ineffective, but because competitive pressures will prevent them from redirecting a meaningful fraction of their AI labor away from capabilities research and towards safety work when it matters most.

The plan to use AI to solve its own safety risks has a critical failure mode: an unlucky ordering of capabilities. If AI becomes a savant at accelerating its own R&D long before it becomes useful for complex tasks like alignment research or policy design, we could be locked into a rapid, uncontrollable takeoff.

The long-held belief that direct human oversight can solve AI risks is breaking down. With sophisticated and dynamic systems, especially agentic ones, a human cannot meaningfully monitor operations in real-time. The solution is shifting towards automated, AI-driven governance and monitoring at higher levels of abstraction.

According to IBM, the key barrier preventing agentic AI systems from moving from impressive demos to widespread production is not a lack of technical capability. The real challenge is the absence of appropriate governance structures and operating models needed to scale these systems safely and effectively.

The very governance bodies created to foster innovation, like AI councils, frequently stifle growth. As projects move from pilot to scale, these groups can become bottlenecks, multiplying reviews and killing momentum because they were designed for permission to start, not permission to grow.

The AI competition is not a race to develop the most powerful technology, but a race to see which nation is better at steering and governing that power. Developing an uncontrollable 'AI bazooka' first is not a win; true advantage comes from creating systems that strengthen, rather than weaken, one's own society.

Unlike conservative data governance focused on protection, AI governance is driven by the race for competitive advantage. Its purpose is less about locking things down and more about enabling the business to "get the rockets off the ground" as quickly and safely as possible, making it a crucial enabler of innovation.

The mismatch between exponentially advancing AI and slow, "medieval" institutions is a core risk. Instead of only focusing on recursively self-improving AI, we should apply technology to create self-improving governance systems that can adapt and update at the same speed as the challenges they face.

A key failure mode for using AI to solve AI safety is an 'unlucky' development path where models become superhuman at accelerating AI R&D before becoming proficient at safety research or other defensive tasks. This could create a period where we know an intelligence explosion is imminent but are powerless to use the precursor AIs to prepare for it.