Recent dramatic advances in artificial intelligence indicate that in the coming years, humanity may irreversibly cross a threshold by creating superhuman general-purpose AI: AI that is better than humans at cognitive tasks in general in the way that AI is currently unbeatable in certain domains. This would upend core aspects of human society, present many unprecedented risks, and is likely to be uncontrollable in several senses. We can choose to not do so, starting by instituting hard limits - placed at the national and international level, and verified by hardware security measures - on the computation that can be used to train and run neural networks. With these limits in place, AI research and industry can focus on making both narrow and general-purpose AI that humans can understand and control, and from which we can reap enormous benefit.

翻译：近期人工智能领域的突破性进展表明，未来数年内人类可能通过创造超人通用人工智能而不可逆转地跨越某个门槛——即在通用认知任务上超越人类表现的人工智能，正如当前AI已在特定领域实现不可战胜。这将颠覆人类社会核心基础，引发诸多前所未有的风险，且可能在多重意义上失去控制。我们可选择阻止这一进程，首先在国家和国际层面设立硬性上限，并通过硬件安全措施加以验证，严格限制可被用于训练和运行神经网络的算力。在此限制下，人工智能研究与产业可聚焦于开发人类能够理解、可控且能带来巨大效益的专用及通用人工智能。