So How Exactly CAN AI Kill Us All? Experts Explain What That Viral Anthropic Tweet Is Referring To

The resignation of Jacob Coxon from Anthropic, following his tenure at OpenAI, underscores a serious internal debate within leading AI labs about the existential risks of superintelligence. While Coxon claims OpenAI employees may not fully grasp the "civilization stakes," he notes that Anthropic employees are well aware but are locked in a competitive race. This internal dissent, particularly from researchers like Nate Soares and Evan Hubinger, who openly state their belief in a significant chance of human extinction from AI, points to a fundamental tension. The drive for technological advancement is outpacing the development of robust safety and alignment protocols. For Asia, this internal struggle within US-based AI giants like Anthropic and OpenAI has direct implications. As Asian nations and companies invest heavily in AI development, they often rely on foundational models and research from these very labs. The lack of a clear alignment plan for superintelligence from these global leaders means that any AI systems built upon their technologies, or developed in parallel by Asian firms, could inherit or replicate these unmitigated risks. The focus for Asian policymakers and tech leaders must shift from merely adopting or developing AI to rigorously integrating safety and ethical frameworks from the outset, rather than assuming global leaders have solved these issues.