美国总统唐纳·川普近期强烈反对对人工智慧进行监管，并将放缓研发以维护安全的呼吁斥为伤害美国资料中心并图利中国的「病态阴谋」，坚称自己就是 AI 所需的唯一护栏；然而专栏作者指出，这种忽视全球协调重要性的态度，恐怕会白白浪费即将登场的美中元首峰会所带来的关键治理契机。

尽管美中双方互信基础薄弱，专家仍认为两国可在共同利益领域逐步取得进展，例如划定禁止使用先进 AI 攻击关键基础设施或核武指挥控制系统的「红线」，并针对中国实验室利用美国模型进行「模型蒸馏」的争议展开协商，甚至透过放宽半导体出口管制等筹码换取对蒸馏行为的约束，以巩固美国硬体市场的主导地位。

面对自主学习 AI 可能失控的潜在风险，美中两国在防范不可控威胁上其实存在共通关切，若双方能推动如第三方独立评估等透明度机制，将有助于建立全球性的安全框架；专家建议若能让川普意识到在 AI 安全领域达成重大外交协议有助于角逐诺贝尔和平奖，或许能促使本次双边会谈取得更加正面的实质突破。

President Donald Trump has firmly pushed back against artificial intelligence regulations, branding calls from industry leaders to slow development for safety reasons as a "SICK conspiracy" favoring China and claiming himself to be the only guardrail AI needs; however, the columnist argues that this stance risks squandering a vital opportunity at the upcoming US-China summit to initiate crucial global safety efforts.

Even with deep bilateral distrust, experts suggest that Washington and Beijing can begin chipping away at risks by targeting areas of mutual and urgent concern, such as establishing redlines against deploying advanced AI to attack critical infrastructure or nuclear command systems, while also addressing the controversial practice of AI distillation with potential trade-offs involving semiconductor export controls.

Shared apprehension regarding self-training AI breaking free of human control provides sensible common ground for establishing global safety frameworks and reciprocal independent model evaluations, and analysts note that if Trump can be persuaded that securing a landmark AI safety agreement with Xi Jinping could pave the way for a Nobel Peace Prize, the talks might yield meaningful progress.