人工智能聊天机器人对科研成果与人才评价的影响研究——基于ChatGPT、Microsoft Bing视角分析
作者:
作者单位:

作者简介:

通讯作者:

中图分类号:

TP18;TP242;G40-058.1

基金项目:


Research on the Influence of AI Chatbot on the Evaluation of Scientific Research Achievements and Talent Evaluation:Analysis based on ChatGPT and Microsoft Bing
Author:
Affiliation:

Fund Project:

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
  • |
  • 文章评论
    摘要:

    美国开放人工智能研究中心(OpenAI)研发的人工智能聊天机器人ChatGPT,基于"生成预训练转换器"(GPT)语言模型能够理解和响应自然语言输入,于2022年11月30日正式发布,短时间内在全球互联网界、学术界及产业界引发热议和关注。以ChatGPT、Microsoft Bing为代表的人工智能聊天机器人基于超大海量数据,采用"从人类反馈中强化学习"(RLHF)进行训练,主要具有三大功能:智能交互对话、智能文本生成、智能文献爬取。人工智能聊天机器人的核心竞争力是采用人工智能生成内容技术(AIGC)而具有强大的内容生成或文本生成能力,除能够生成一般应用性文稿外,还能够从事科学写作,从而引领和推动知识生产方式和学术研究范式变革,知识生产和学术研究进程将显著提速,知识生产者将从非创造性劳动中获得解放,知识生产方式将由"手动式""半自动式"迈向"自动化""智能化"。这种变革对传统的科研成果评价机制将产生重大影响,更进一步将对通过科研成果进行人才评价产生重大影响。大体上说,以ChatGPT为代表的人工智能聊天机器人应用于学术研究中,将遮蔽成果演化学术史,影响科研成果创新性评价;挑战知识产权保护制度,影响科研成果权属认定;带来学术伦理风险,影响科研成果评价生态;冲击人才评价基本标准,影响人才评价客观性。鉴于人工智能的渗透性、扩散性和颠覆性,首先要在思想认识上积极面对而不是消极应对,用其所长、避其所短,审慎使用,着力发挥其促进知识生产、推动学术进步的积极作用;其次,要抓住现行相关法规制度与快速进步的人工智能技术之间发展不平衡的矛盾,进一步完善制度、设置"护栏",规范使用;第三,要推动技术开源,公开其智能生成逻辑,增强人工智能聊天机器人的技术可信度;第四,要建立人机协作机制,既发挥技术的数据处理优势,又发挥人的理性思维优势,最终由人把关知识生产、科研成果及人才评价。

    Abstract:

    Chat GPT, an artificial intelligence (AI) chatbot developed by the United States Open Artificial Intelligence Research Center (OpenAI), can understand and respond to natural language input based on the language model of "Generate Pre-training Converter" (GPT), which was officially released on November 30, 2022, and has caused heated discussion and attention in the global Internet, academia and industry in a short time. The AI chatbots represented by ChatGPT and Microsoft Bing are based on huge amounts of data and trained with "reinforcement learning from human feedback" (RLHF). They have three main functions: intelligent interactive dialogue, intelligent content generation, and intelligent document crawling.The core competitiveness of the AI chatbot is to adopt the AI generated content technology (AIGC) and have strong content generation or text generation ability. In addition to generating general applied manuscripts, it can also engage in scientific writing, thus leading and promoting the transformation of the mode of knowledge production and academic research paradigm. The process of knowledge production and academic research will be significantly accelerated, and knowledge producers will be liberated from non-creative labor, the mode of knowledge production will move from "manual" and "semi-automatic" to "automatic" and "intelligent".This change will have a significant impact on the traditional evaluation mechanism of scientific research achievements, and further will have a significant impact on talent evaluation. In general, the application of AI chatbot represented by ChatGPT in academic research will obscure the evolution of academic history of achievements and affect the innovative evaluation of scientific research achievements, and challenge the intellectual property protection system and affect the ownership of scientific research achievements, and bring academic ethical risks and affect the evaluation ecology of scientific research achievements, and impact the basic standards of talent evaluation and affect the objectivity of talent evaluation. In view of the penetration, diffusion and subversion of AI, first of all, we should face it positively rather than passively in ideological understanding, utilize its strengths and avoid its weaknesses prudently, and exert its positive role in promoting knowledge production and promoting academic progress. Secondly, institution construction should be further improved to set up "guardrails" for AI chatbot according to the imbalance between the current relevant laws and regulations and the rapid progress of AI technology. Third, AI chatbot technology open source should be promoted to disclose its intelligent generation logic, and enhance the technical credibility of AI chat robots. Fourth, human-computer cooperation mechanism should be established to give play to the advantages of technology in data processing and human rational thinking, and finally let people check the knowledge production, scientific research achievements and talent evaluation.

    参考文献
    相似文献
    引证文献
引用本文

蒋华林.人工智能聊天机器人对科研成果与人才评价的影响研究——基于ChatGPT、Microsoft Bing视角分析[J].重庆大学学报社会科学版,2023,(2):97-110. DOI:10.11835/j. issn.1008-5831. pj.2023.02.003

复制
分享
文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 收稿日期:
  • 最后修改日期:
  • 录用日期:
  • 在线发布日期: 2023-05-08
  • 出版日期: