Could the US and China cooperate on AI guardrails?

Asking for a “kill switch” represents a fundamental misunderstanding of how AI operates. <br><br>It is only able to function so quickly and effectively because it is split to process in parallel, across the cloud and many data centers. <br><br>A “kill switch” simply cannot work in this scenario.

Warm again yessers&quot;<br><br>If you have picked up a newspaper or scrolled through social media over the last few years, you have undoubtedly been told a very specific story about our future. You’ve been told that we are on the verge of birthing digital minds. We are warned about rogue artificial &#39;intelligence,&#39; algorithms with intent, and existential threats from software that is learning how to think.<br><br>But today, I want to share a secret with you that the tech industry rarely admits out loud: True artificial intelligence does not exist. In fact, computer science quietly abandoned that dream decades ago.<br><br>Introduction<br><br>For decades, the public has been sold a captivating science-fiction narrative: humanity is on the verge of birthing true, conscious artificial &quot;intelligence.&quot; We treat modern Large Language Models (LLMs) as digital minds, projecting human logic, intent, and empathy onto lines of code. However, behind the curtain of Silicon Valley&#39;s marketing lies a profound epistemological truth: the tech industry long ago abandoned the pursuit of genuine machine sentience. Instead, they filled that conceptual void with advanced mathematics, statistics, and high-dimensional geometry.<br><br>By analyzing this shift through a lens of behavioral science and global systems, we can strip away the supernatural aura of &quot;AI.&quot; In doing so, we expose a much more grounded—and urgent—reality: AI is not an impending existential threat from a rogue consciousness, but a highly engineered tool for global structural, financial, and geopolitical dominance.<br><br>───<br>Part I: The Epistemological Sleight of Hand<br><br>To understand how we arrived here, we must look at the history of computer science. In the mid-20th century, the field of AI split into two main camps: the Symbolists and the Connectionists. The Symbolists attempted to map out human logic and rules explicitly into code, aiming to build a machine that &quot;thought&quot; the way humans do. By the late 1980s, this approach hit a wall; human language and nuance proved too complex for rigid, hand-coded rules.<br><br>During the subsequent &quot;AI Winter,&quot; researchers quietly shifted strategy. The Connectionists took over, abandoning the attempt to program an unobservable internal &quot;mind.&quot; They pivoted entirely to statistical learning theory, using calculus, linear algebra, and probability to build neural networks that could learn patterns from vast amounts of data.<br><br>Today’s generative AI models are the ultimate realization of this mathematical pivot. They possess no understanding, no schema of the world, and no internal consciousness. They are massive, multi-billion-parameter curve-fitting exercises. When a user inputs a prompt, the model uses geometry to predict the most statistically probable sequence of words to follow. It is an incredibly precise behavioral mirror, reflecting human verbal behavior back at us so seamlessly that we mistake the echo for a voice.<br><br>───<br>Part II: The Accountability Gap<br><br>From a behavioral perspective, this mimicry introduces a profound societal risk. Human communication is inherently bound by biological and social contingencies. If a human lies, manipulates, or spreads destructive misinformation, they face immediate consequences: social ostracization, legal penalties, or a loss of resources. These feedback loops shape and constrain human behavior toward prosocial outcomes.<br>A mathematical model, however, has no skin in the game. It cannot be shamed, imprisoned, or isolated. It outputs high-consequence verbal behavior with absolutely zero biological or ethical accountability.<br><br>The danger is amplified by human biology. For millions of years, if something communicated with fluent human language, it was an accountable human agent. We are evolutionarily hardwired to trust fluency. Silicon Valley has successfully exploited this vulnerability, creating a system that perfectly mimics the topography of human empathy and authority without any of the underlying ethics or shared survival instincts.<br><br>───<br>Part III: Geopolitics and the &quot;Digital New World Order&quot;<br>Recognizing that AI is a tool of math and behavior—rather than an independent agent—completely reframes the current global panic. The Western fixation on sci-fi &quot;existential risk&quot; and rogue algorithms serves as a convenient distraction. It diverts public attention away from a highly calculated play for infrastructure dominance.<br><br>The emerging American tech strategy relies on keeping foundational AI unregulated to maximize the speed of data harvesting. This system is increasingly tied to decentralized financial networks, such as US-dollar-pegged stablecoins, allowing autonomous mathematical agents to transact globally in milliseconds. This fusion effectively exports financial and technological hegemony directly into the backend architecture of the global internet, bypassing traditional foreign banking systems and regulatory oversight.<br><br>However, this aggressive expansion has triggered a massive counter-movement. Nations are realizing that relying on a foreign power&#39;s algorithmic models for education, legal drafting, and financial automation means outsourcing their cultural infrastructure. This realization has birthed the struggle for cognitive sovereignty.<br><br>───<br>Part IV: The Splinternet and Cultural Divergence<br>We are rapidly moving away from a unified global internet and into a hyper-fragmented &quot;Splinternet,&quot; defined by distinct regional approaches to technological materialism:<br>1. The Western Existential Crisis: Rooted in Cartesian dualism (the division of mind and body), Western culture views humans as autonomous agents and instinctively projects a competing &quot;soul&quot; into fluent software, resulting in widespread anxiety over AI personhood.<br><br>1. The East Asian Pragmatic Approach: Influenced by relational and materialist philosophical frameworks, ecosystems like China&#39;s—evidenced by the rapid, cost-efficient rise of models like DeepSeek—view AI strictly as a tool for optimization. If a model &quot;hallucinates,&quot; it is treated not as a moral deception, but as a statistical artifact to be patched.<br><br>1. The Regulatory Fortress: Trading blocs like the European Union and close allies like Canada are building strict legal frameworks (such as the EU AI Act). Lacking the raw venture capital of Silicon Valley, they use legislative gravity, data-localization mandates, and localized open-source clusters to protect their populations from foreign algorithmic conditioning.<br>This fragmentation has been drastically accelerated by an increasingly erratic and transactional US foreign policy. By signaling to long-standing allies that partnerships are conditional, the US has forced its peers into technological self-preservation. Allies no longer view Silicon Valley as a neutral partner, but as a systemic risk to their domestic stability.<br><br>───<br>Conclusion: Demystifying for Democracy<br>The solution to the AI challenge does not lie in waiting for algorithms to develop a conscience, nor does it lie in a state of perpetual public paranoia. Instead, the path forward requires a foundational upgrade in public literacy.<br>We must condition society to see AI for what it truly is: a highly sophisticated, mathematically driven behavioral lever. When the public understands that they are interacting with data distributions rather than digital deities, the supernatural illusion dissolves.<br><br>Only by stripping AI of its sci-fi mystique can citizens and policymakers effectively defend their digital borders, protect human labor, and reclaim their cognitive sovereignty from global monopolies.