A classification of the global catastrophic risks of AI is presented, along with a comprehensive list of previously identified risks. This classification allows the identification of several new risks. We show that at each level of AI’s intelligence power, separate types of possible catastrophes dominate. Our classification demonstrates that the field of AI risks is diverse, and includes many scenarios beyond the commonly discussed cases of a paperclip maximizer or robot-caused unemployment. Global catastrophic failure could happen at various levels of (...) AI development, namely, before it starts self-improvement, during its takeoff, when it uses various instruments to escape its initial confinement, or after it successfully takes over the world and starts to implement its goal system, which could be plainly unaligned, or feature-flawed friendliness. AI could also halt at later stages of its development either due to technical glitches or ontological problems. Overall, we identified around several dozen scenarios of AI-driven global catastrophe. The extent of this list illustrates that there is no one simple solution to the problem of AI safety, and that AI safety theory is complex and must be customized for each AI development level. (shrink)

Many global catastrophic risks are threatening human civilization, and a number of ideas have been suggested for preventing or surviving them. However, if these interventions fail, society could preserve information about the human race and human DNA samples in the hopes that the next civilization on Earth will be able to reconstruct Homo sapiens and our culture. This requires information preservation of an order of magnitude of 100 million years, a little-explored topic thus far. It is important that a potential (...) future civilization will discover this information as early as possible, thus a beacon should accompany the message in order to increase visibility. The message should ideally contain information about how humanity was destroyed, perhaps including a continuous recording until the end. This could help the potential future civilization to survive. The best place for long-term data storage is under the surface of the Moon, with the beacon constructed as a complex geometric figure drawn by small craters or trenches around a central point. There are several cost-effective options for sending the message as opportunistic payloads on different planned landers. (shrink)

Pandemics have been suggested as global risks many times, but it has been shown that the probability of human extinction due to one pandemic is small, as it will not be able to affect and kill all people, but likely only half, even in the worst cases. Assuming that the probability of the worst pandemic to kill a person is 0.5, and assuming linear interaction between different pandemics, 30 strong pandemics running simultaneously will kill everyone. Such situations cannot happen naturally, (...) but because biotechnology is developing analogously to Moore’s law, it may become possible in the near future (10-50 years from now), because of biohackers, CRISPR, bioprinters, AI-assisted DNA-programing, and weapons of “knowledge-enabled mass destruction” published on the Internet. It could also happened in case of large-scale biological war, or if a rogue country released its entire biological arsenal simultaneously. We also will address other scenarios and risk increasing factors as well as mitigation and adaptation strategies. (shrink)

Abstract: The field of artificial general intelligence (AGI) safety is quickly growing. However, the nature of human values, with which future AGI should be aligned, is underdefined. Different AGI safety researchers have suggested different theories about the nature of human values, but there are contradictions. This article presents an overview of what AGI safety researchers have written about the nature of human values, up to the beginning of 2019. 21 authors were overviewed, and some of them have several theories. A (...) theory classification method is suggested, where the theories are judged according to the level of their complexity and behaviorists-internalists scale, as well as the level of their generality-humanity. We suggest that a multiplicity of well-supported theories means that the nature of human values is difficult to define, and some meta-level theory is needed. (shrink)