Jacob Coxon, a former core technical researcher at Anthropic and OpenAI, resigned in late September 2026 and issued a stern warning regarding the development of artificial intelligence (AI). Coxon alleged that a "doomsday mentality" is prevalent among practitioners within cutting-edge laboratories, with many genuinely believing that AI could lead to human extinction before 2030. This event occurs against the backdrop of an intensifying artificial intelligence race, as major tech giants invest massive amounts of capital and computing resources in hopes of gaining a lead in the development of superintelligence.

Meanwhile, the geopolitical landscape remains volatile. According to a report by Liveuamap on September 21, 2026, a drone attack launched by Ukraine on September 20 resulted in the disruption of operations at relevant energy facilities, highlighting the vulnerability of global energy infrastructure amidst conflict.

Multiple media sources have confirmed that there is a serious divide within the tech industry regarding risk assessments for the future of AI. Coxon’s resignation announcement explicitly stated that tech giants are wagering the lives of all humanity on the superintelligence race, ignoring potential catastrophic risks. The tech community’s reaction to this allegation has been polarized; a camp led by the open-source platform Hugging Face has expressed strong skepticism, arguing that current concerns regarding AI doomsday theories lack empirical evidence. They have even likened the discourse to "air conditioning repairmen talking about climate change," implying that some engineers may be exaggerating the social impact of technological development while lacking a comprehensive understanding of macro-systems.

However, significant differences remain in how various sources assess AI risks, and some information remains to be clarified. Coxon’s argument focuses on the psychological expectations and potential technical loss of control among internal practitioners, yet he has not proposed a specific path to technical destruction. Conversely, the rebuttal from the CEO of Hugging Face emphasizes an alternative perspective from the open-source community on technology governance, suggesting that fears surrounding closed-source development may actually hinder transparency and safety.

Furthermore, regarding the information environment in East Asia, a report by Voice of America (VOA) noted that the distributor for Reuters had faced information filtering issues, including news related to protests in Hong Kong; even some reports from China’s official news agency, Xinhua, had been blocked. Such restrictions on the dissemination of information lead to vastly different perceptions of AI policy and geopolitical conflicts across different regions.

These events have had tangible impacts on global energy, industrial policy, and social stability. First, the Ukrainian drone attacks on energy facilities directly threaten the stability of transnational energy supply chains and may indirectly affect price volatility for energy-importing nations in East Asia. Second, the controversy over the safety of AI research has forced governments worldwide to re-examine their regulatory frameworks for artificial intelligence. If tech giants fail to effectively address concerns from their own engineers regarding "AI spiraling out of control," it may trigger stricter administrative interventions, which would in turn affect the pace of R&D and capital flow in the global AI industry.

Moreover, restricted information flow, particularly in East Asia, not only affects the public's right to know about major public issues but also deepens the divide within the international community regarding technology governance and political transparency.

Despite the heated discussion surrounding AI doomsday theories, there are clear limitations to the perspectives involved, and they should not be over-interpreted. Coxon’s view reflects the anxiety of some top researchers, but this does not mean that AI will inevitably lead to destruction, nor can it serve as the sole basis for dismissing technological progress. Similarly, while the analogy of air conditioning repairmen discussing climate change is ironic, it may also overlook the critical role that technology developers play in system architecture.

Additionally, the attacks on energy facilities during geopolitical conflicts are the result of specific military actions. While they share the category of global security issues with AI development risks, there is no direct causal link between the two, and observers should avoid conflating them to prevent misjudgments of the global situation.

Future observations should focus on three aspects: First, whether cutting-edge AI laboratories will adjust their R&D transparency policies in response to internal waves of resignations and public opinion pressure; second, whether international coordination mechanisms for the protection of energy infrastructure can effectively address the threats posed by drone attacks; and third, how East Asia will adjust its digital governance strategy in the face of dual pressures from information blockades and technological competition. These issues will continue to influence the path of global technological development and geopolitical stability in the coming years, and are worthy of close attention.