A Moral Imperative: The Need for Continual Superalignment of Large Language Models

arXiv:2403.14683v1 Announce Type: cross
Abstract: This paper examines the challenges associated with achieving life-long superalignment in AI systems, particularly large language models (LLMs). Superalignment is a theoretical framework that aspires to ensure that superintelligent AI systems act in accordance with human values and goals. Despite its promising vision, we argue that achieving superalignment requires substantial changes in the current LLM architectures due to their inherent limitations in comprehending and adapting to the dynamic nature of these human ethics and evolving …

abstract act ai systems arxiv challenges continual cs.ai cs.cl cs.cy cs.lg framework human language language models large language large language models life llms paper superalignment superintelligent ai systems type values vision