Good to see China pushing back against a militaristic, winner-takes-all framing of AI — escalating into a geopolitical sprint is exactly what will make the technology more dangerous, not safer. But rejecting “threat narratives” doesn’t let Beijing off the hook: any global framework must insist on human rights, press freedom, independent oversight and limits on surveillance and weaponization, not just state control dressed up as “security.” Calls from industry to slow down development are sensible, but a slowdown that’s meaningful has to be multilateral, transparent, verifiable and include civil society, labor representatives and nations of the Global South — it can’t be a backroom deal to freeze one country out or cement corporate power. Equally worrying is the Trump-era “whoever wins AI wins” mindset: competition without guardrails will deepen inequality, entrench surveillance capitalism, and put lives at risk. Progressives should push for binding international rules, open auditability of powerful models, strong labor protections, whistleblower protections for safety researchers, and enforceable bans on offensive military uses — because AI governance must center people, rights and democracy, not just national prestige or corporate profit.