Former OpenAI chief scientist Ilya Sutskever reportedly suggested building a 'doomsday bunker' to safeguard key researchers from potential disasters linked to artificial general intelligence (AGI). Sutskever, known for his work on ChatGPT, voiced his concerns during a 2023 meeting, stating the bunker would be built before AGI's release. This proposal highlights growing anxieties within the AI community regarding AGI's unpredictable nature and potential risks to humanity, with some experts fearing a 'rapture' scenario.
The bunker idea reflects broader discussions about AI safety and preparedness as AI systems become more powerful. Sutskever's concerns, detailed in Karen Hao's upcoming book 'Empire of AI,' underscore the need for caution in AGI development. Other AI leaders, like DeepMind CEO Demis Hassabis, have also expressed worries about society's readiness for AGI, raising questions about the ethical and practical implications of rapidly advancing AI technology. The concept of AGI, which refers to AI surpassing human cognitive abilities, remains a hotly debated topic, with experts divided on its potential benefits and existential threats.
