The current bottleneck is political will, not research

Short summary

The author argues that AI safety's main bottleneck is political will, not research: best practices exist but aren't applied because most influential policymakers have never had a serious conversation about catastrophic AI risk. The field over-invests in research relative to advocacy (~3.6 researchers per advocate), and industry out-lobbies civil society 7:1 at the EU level. The post calls for shifting effort toward moving policymakers through an awareness-to-champion pipeline.

• •AI safety bottleneck is political will, not lack of research or policy ideas
• •Majority of top policymakers have never had a serious conversation about AI x-risk
• •Field invests ~3.6x more in research than advocacy; industry out-lobbies civil society 7:1 at EU level
• •Additional effort should prioritize advocacy and engagement over producing more papers

Generated with AI, which can make mistakes.