## "Safety" or "Control"? a critical linguistic sleight-of-hand in today’s AI discourse: safety as a trojan horse to justify hoarding power (a classic political maneuver to consolidate authority.) When politicians, commentators, and even some tech leaders talk about ensuring AI "safety," what they are really asking for is **centralized control**. The narrative of "safety" is used as a Trojan horse to justify hoarding power, restricting access to compute, and dictating what AI models are allowed to learn or output ## The Eerie Parallels Between Controlling AI and Controlling Humans The arguments for centralized control of Als and of people are eerily similar. They are based in fear. They are all about us vs them. They demonize the other. They claim the other can't be trusted. the underlying political philosophy for both is identical -- **They are both based on fear.** They both stem from a desire by a centralized authority to manage entities (whether human populations or artificial minds) that they feel they cannot fully predict or trust. ## The Fallacy of the "Us vs. Them" Demonization The politics of fear always requires an "Other" to demonize. In human politics, this is often xenophobia or nationalism (fearing other countries or cultures). In the current tech landscape, **AI has become the ultimate "Other."** Sutton notes that fear-mongers frame AI as an alien, untrustworthy entity that is out to replace or destroy us. By demonizing AI, politicians and regulators justify their calls for strict control. [[AI is fundamentally us]] ## The Core Solution: Decentralized Cooperation vs. Centralized Control All human flourishing—everything good in the world, including the economy, scientific exchange, and peaceful societies—comes from decentralized cooperation. Humans interact, trade, and work together voluntarily. While humans can be terrible at cooperation (leading to wars), it remains the ultimate source of progress. **Human-AI flourishing must follow the exact same model**. Instead of placing a global, centralized regulatory "lid" on AI development, we should integrate AI into our society through decentralized networks of cooperation. We should treat intelligent machines as new participants in our global cooperative economy, rather than as prisoners that need to be chained and monitored by a central committee. ## A Call to Resist Authoritarian Instincts When faced with rapid change and a new form of intelligence, humanity's knee-jerk reaction is mistrust, non-cooperation, and a demand for a strong central authority to "protect" us. **"look with open eyes"** at the people and organizations calling for this centralized control and resist those calls