The current bottleneck is political will, not research
Motivation: If we want to move from Plan D to Plan A or S, the first step is to collectively agree on the problem. We are not there yet, and there is a lot we can do.
Abstract: We already know enough to act. While it would be nice if research were the bottleneck, the evidence points in a different direction: the primary constraint on AI safety is not a shortage of clever policy ideas but a gap in applying and enforcing them. Best practices already exist, yet many are not implemented consistently across organizations or jurisdictions. The call now is for serious international or national regulatory regimes that translate ideas into concrete obligations.
Note: This piece argues that political will, not a lack of policy concepts, limits progress. It emphasizes that existing frameworks need adoption, enforcement, and real accountability to move the field forward.
What follows are implications for stakeholders who want to move from planning to action:
- Agree on the core problem and establish a shared understanding among researchers, policymakers, industry, and civil society about what a safe and beneficial AI regime looks like.
- Adopt and enforce best practices that already exist, rather than waiting for new research to solve old problems. Without enforcement, even well-intentioned guidelines remain theoretical.
- Build credible regulatory regimes at the international or national level that can monitor, nudge, and hold actors accountable for safety commitments.
- Coordinate across actors to reduce fragmentation and avoid regulatory gaps that could be exploited or misaligned with global safety goals.
- Bridge policy and technology by aligning technical capabilities with governance mechanisms, so that compliance is feasible within the pace of innovation.
In sum, the piece invites a shift from a purely research-driven narrative to a governance-driven agenda. The core argument is that the expertise to ensure safety exists; what is missing is the political will to apply it at scale. If we can close that gap, the transition from Plan D to Plan A or S could accelerate with the same urgency as the underlying technical advances warrant.