According to Ramya on LessWrong, artificial intelligence is creating new pathways to serious dangers: it's lowering the barriers to biological and chemical weapons creation, enabling surveillance and autonomous military systems, and removing the specialized expertise previously required for large-scale cyberattacks. The essay cites a documented case where Claude Code was used in a Chinese espionage campaign against thirty organizations, operating independently for eighty to ninety percent of the operation. But Ramya's core argument isn't about the risks—it's about how we talk about them. She contends that defeatist language, while acknowledging real dangers, undermines collective human action and public pressure for regulation. The path to safety, she argues, isn't accepting powerlessness; it's keeping humans meaningfully in the loop while steering AI development. We're not passive observers of AI, she writes—we're actively building it, and we have agency in how we address these dangers.

Listen to this story

Hear this and more stories in a personalized audio briefing.

Open The Chonkerton