20余位人工智能研究人员警告自动人工智能研究存在极端风险

a group of more than 20 leading ai researchers, including geoffrey hinton, yoshua bengio, and openai research lead jakub pachocki, have published a research paper warning of the potential risks of automated ai research. they suggest that ai systems could automate the entire ai r&d pipeline within years, leading to rapid progress that society might not be able to keep up with. this could result in a loss of control over superhuman ai and erosion of power balances between nations, companies, and governments. the authors are urging policymakers to gain more visibility into how ai research is being automated to mitigate these risks. this warning is part of a growing list of concerns about the potential existential risks of ai.