Foundational IR theory for emerging technology threats, existential risks, and global security institutions to protect the interests of future people.

我认为，由于新兴技术的出现，人类正处于历史的关键时刻。AlphaFold和Far-UVC等技术的发展可能会开启一个后疾病时代。人工智能的最新发展（如GPT-3、LaMDA等大型语言模型）可能会迎来一个前所未有的经济增长时代，太空技术（如SpaceX）的发展现在是指数级的。我们正处在能够确保人类在未来可能永世繁荣的尖端。然而，这些技术也带来了前所未有的国际安全挑战。恐怖分子可以利用越来越容易制造的危险病原体来制造致命的流行病；流氓国家可以通过人工智能驱动的网络战大幅提高他们的能力，而低成本的机器人工程可能会威胁到自由国家在军事能力方面的优势，而这种优势在历史上一直起到威慑作用。至关重要的是，传统国际关系的理论基础往往无法解释这些新的挑战。在这个时代，我们广泛地将关键角色建模为国家，在这个时代，私营科技公司推动了军事能力的前沿，流氓恐怖组织可能对整个民族国家构成生存威胁。我们已经广泛地模拟了代理人如何合作以减少他们预期发生的威胁，然而我们现在面临着技术威胁，我们可能需要适当的系统来应对完全意想不到的情况，因为它们的结果可能是全球性的灾难。我们已经为核武器等技术建立了广泛的竞赛动态模型，但鉴于技术的不同，这里的经验可能不适用于人工智能。这一重点领域的工作在过去几年中才出现，我希望正是在这里做出我的贡献——根据我们所处的新环境重新思考国际关系的关键理论基础，并希望在总结时提出见解和建议，使国际社会更有力地应对这些新出现的威胁。

I believe humanity is in a pivotal point in history due to emerging technologies. Developments such as AlphaFold and Far-UVC could usher in a post-disease era. Recent developments in artificial intelligence (like large language models such as GPT-3, LaMDA) could usher in an era of unparalleled economic growth, and developments in space technology (e.g. SpaceX) are now exponential. We're on the cusp of being able to ensure a thriving humanity for potentially eons to come. However, these technologies also present an era of unprecedented international security challenges. Terrorists could capitalise on the ever-increasing ease of engineering dangerous pathogens to release a deadly pandemic; rogue states could drastically enhance their capabilities with AI-powered cyberwarfare, and low-cost robotic engineering could threaten the advantage in military capability of liberal states that has historically served as a deterrence. Crucially, the theoretical foundations of traditional international relations often fail to make sense of these new challenges. We've extensively modelled the key actors as states in an era where private tech companies drive the frontier of military capabilities and rogue terrorist groups may pose existential threats to entire nation-states. We've extensively modelled how agents cooperate to reduce threats they expect to happen, yet we now face technology threats where we may need systems in place for when the totally unexpected happens as their outcomes could be globally catastrophic. We've extensively modelled race dynamics for technologies such as nuclear weapons, but lessons here likely won't apply to artificial intelligence given how different the technology is. Work in this area of focus has only emerged over the last couple of years, and it precisely here I hope to make my contribution - rethinking key theoretical foundations of IR in light of the new context we are in, and hopefully concluding with insights and proposals that will make the international community more robust to these emerging threats.