Scholarlink

<p>Large Language Models (LLMs) have emerged as a major innovation in the field of AI due to their remarkable ability to handle diverse tasks. However, people remain divided on whether LLMs can ultimately achieve Artificial General Intelligence (AGI). Proponents argue that the versatility demonstrated by LLMs is a significant step towards AGI; critics, on the other hand, contend that LLMs merely simulate the surface features of language without possessing the attributes of human intelligence and cannot be AGI. The former group is overly optimistic, believing that challenges such as hallucinations and reasoning will eventually be resolved, while the latter group typically holds an anthropocentric view, asserting that AGI needs to simulate human intelligence to some extent. That`s supported by the high standards for AGI, though it may be ad hoc. The debate stems from differing interpretations of AGI, and the gap can be bridged through a new technology called Large Language Models for Games. This approach leverages the strengths of LLMs while aligning with traditional AGI research pathways, addressing the concerns of skeptics and shifting the focus to the true core of AGI: the ability to handle novel problems in unfamiliar scenarios.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

The Potential of Large Language Models to Achieve Artificial General Intelligence

  • 期刊论文

  • 发表日期：

    2025年6月4日

查看全文
  • Hao Zhan

摘要

Large Language Models (LLMs) have emerged as a major innovation in the field of AI due to their remarkable ability to handle diverse tasks. However, people remain divided on whether LLMs can ultimately achieve Artificial General Intelligence (AGI). Proponents argue that the versatility demonstrated by LLMs is a significant step towards AGI; critics, on the other hand, contend that LLMs merely simulate the surface features of language without possessing the attributes of human intelligence and cannot be AGI. The former group is overly optimistic, believing that challenges such as hallucinations and reasoning will eventually be resolved, while the latter group typically holds an anthropocentric view, asserting that AGI needs to simulate human intelligence to some extent. That`s supported by the high standards for AGI, though it may be ad hoc. The debate stems from differing interpretations of AGI, and the gap can be bridged through a new technology called Large Language Models for Games. This approach leverages the strengths of LLMs while aligning with traditional AGI research pathways, addressing the concerns of skeptics and shifting the focus to the true core of AGI: the ability to handle novel problems in unfamiliar scenarios.