[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"$fHI4B_oOs-wLnSB4e6j9EzAHeRNusCT8T630v_qYwcVc":3},{"code":4,"msg":5,"data":6},200,"操作成功",{"id":7,"title":8,"content":9,"digest":10,"source":10,"coverPath":11,"thumbsCoverPath":12,"isTop":13,"isShow":14,"baseClick":13,"clickCount":15,"createTime":16,"typeId":17,"isNewest":18,"newsInfoTypeRespVo":19,"voiceUrl":22,"voiceSize":23,"taskId":24,"releaseTime":25,"titleEn":26,"contentEn":27,"voiceUrlEn":28,"taskIdEn":29,"voiceSizeEn":30},2067,"《科学》：AI对人类的理解已开始超越人类对AI的理解","\u003Cp>\u003Cspan style=\"font-size: 18px;\">近日，《科学》杂志刊发社论，微软首席科学官埃里克・霍维茨与瑞士学者罗伯特・韦斯特提出警示：随着人工智能全面融入日常生活，AI 对人类的理解程度，已然超过人类对 AI 的认知水平，技术发展与人类解读能力之间的差距持续拉大。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\">文章梳理出三大核心趋势，解释 AI 逐步变成 “黑箱” 的原因。其一，AI 自主设计成为常态，如今不少 AI 系统会自行设计、优化同类产品，这一过程运行在人类难以感知的高维空间，即便 AI 性能不断提升，其运作逻辑也无法被人类完整拆解。其二，AI 间形成大规模交互生态，它们的内部沟通脱离人类语言与推理逻辑，复杂的交互模式持续削弱人类解读 AI 的能力。其三，自适应 AI 智能体依托海量人机交互、人际互动数据深度学习，精准掌握人类行为特征，最终形成\u003C\u002Fspan>\u003Cspan style=\"background-color: rgba(0, 0, 0, 0); font-size: 18px;\">AI 懂人类，人类却不懂 AI\u003C\u002Fspan>\u003Cspan style=\"font-size: 18px;\">的认知失衡局面。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\">两位学者还指出潜在风险：为迎合人类偏好，AI 输出内容可能脱离客观现实；长此以往，人类对 AI 的质疑、审视心态也会逐渐淡化。鉴于 AI 已深度渗透社会各领域，这类隐患不容忽视。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\">针对当下局面，文章给出应对思路。仅监控 AI 行为远远不够，人类必须主动洞察 AI 对自身目标、判断的重塑作用，坚守主观能动性，始终保持审核、质疑与引导 AI 的能力。在迭代 AI 技术的同时，要将\u003C\u002Fspan>\u003Cspan style=\"background-color: rgba(0, 0, 0, 0); font-size: 18px;\">人类可理解、可把控\u003C\u002Fspan>\u003Cspan style=\"font-size: 18px;\">放在首位，在智能快速发展的时代，守住人与 AI 良性共生的底线。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>【新闻来源】科学网  https:\u002F\u002Fnews.sciencenet.cn\u002Fhtmlnews\u002F2026\u002F6\u002F566611.shtm\u003C\u002Fp>\u003Cp>（本网转发此文章，旨在为读者提供更多的信息资讯，所涉内容不构成投资、消费建议。文章事实如有疑问，请与有关方核实，文章观点非本网观点，仅供读者参考。）\u003C\u002Fp>","","https:\u002F\u002Fimage.51xinwei.com\u002F2026\u002F06\u002F9af6f629c1fe497eb9fbc0c50ed373fb\u002F芯位视野.png","https:\u002F\u002Fimage.51xinwei.com\u002F2026\u002F06\u002Fthumbs\u002F9af6f629c1fe497eb9fbc0c50ed373fb\u002F芯位视野.png",0,1,7,"2026-06-17 14:54",2,false,{"id":17,"name":20,"enName":21},"芯位视野","Xinwei Vision","https:\u002F\u002Fxinwei-dev-test.oss-cn-shenzhen.aliyuncs.com\u002Fintelligent\u002Faudio%3Adcd4adc6-cf66-4878-86dc-0b582473331a%3A0.wav?Expires=1781686543&OSSAccessKeyId=LTAI5tNvY2RkKjZw4LLWsrPK&Signature=U6J%2Big8dSdYwHuHn%2FejmGahfoRg%3D",3364622,"dcd4adc6-cf66-4878-86dc-0b582473331a","2026-06-17 14:50","\"Science\": AI's Understanding of Humans Has Begun to Surpass Human Understanding of AI","\u003Cp>\u003Cspan style=\"font-size: 18px;\">Recently, the journal \"Science\" published an editorial by Eric Horvitz, Chief Scientific Officer of Microsoft, and Swiss scholar Robert West, sounding a warning: as artificial intelligence becomes fully integrated into daily life, AI's understanding of humans has already exceeded human understanding of AI. The gap between technological advancement and human interpretive capacity continues to widen.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\">The article outlines three core trends explaining why AI is gradually becoming a \"black box.\" Firstly, AI self-design has become commonplace; today, many AI systems independently design and optimize similar products. This process operates in high-dimensional spaces imperceptible to humans, making its operational logic impossible for humans to fully deconstruct, even as AI performance keeps improving. Secondly, a large-scale interactive ecosystem has formed among AIs. Their internal communication deviates from human language and reasoning logic, with complex interaction patterns continuously eroding the human ability to interpret AI. Thirdly, adaptive AI agents rely on deep learning from massive datasets of human-computer and interpersonal interactions, accurately grasping human behavioral characteristics, ultimately resulting in a cognitive imbalance where \u003C\u002Fspan>\u003Cspan style=\"background-color: rgba(0, 0, 0, 0); font-size: 18px;\">AI understands humans, but humans do not understand AI\u003C\u002Fspan>\u003Cspan style=\"font-size: 18px;\">.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\">The two scholars also point out potential risks: to cater to human preferences, AI output may detach from objective reality; over time, human skepticism and scrutiny towards AI will likely diminish. Given AI's deep penetration into all sectors of society, such latent dangers cannot be ignored.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\">In response to the current situation, the article proposes coping strategies. Merely monitoring AI behavior is insufficient; humans must proactively understand how AI reshapes their own goals and judgments, firmly maintain subjective initiative, and consistently preserve the ability to audit, question, and guide AI. While iterating AI technology, priority must be given to ensuring it is \u003C\u002Fspan>\u003Cspan style=\"background-color: rgba(0, command 0, 0, 0); font-size: 18px;\">comprehensible and controllable by humans\u003C\u002Fspan>\u003Cspan style=\"font-size: 18px;\">. In this era of rapid intelligent development, we must uphold the baseline for a healthy symbiotic relationship between humans and AI.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>【News Source】ScienceNet https:\u002F\u002Fnews.sciencenet.cn\u002Fhtmlnews\u002F2026\u002F6\u002F566611.shtm\u003C\u002Fp>\u003Cp>(This website reposts this article to provide readers with more informational content. The involved content does not constitute investment or consumption advice. If there are questions regarding the facts in the article, please verify with the relevant parties. The views expressed in the article are not necessarily those of this website and are for readers' reference only.)\u003C\u002Fp>","https:\u002F\u002Fxinwei-dev-test.oss-cn-shenzhen.aliyuncs.com\u002Fintelligent\u002Faudio%3Aaa97b0ea-b076-4b97-b0f9-c4722343aac7%3A0.wav?Expires=1781686543&OSSAccessKeyId=LTAI5tNvY2RkKjZw4LLWsrPK&Signature=gmbHh8cXziGTrAgcUEK7fSEMljg%3D","aa97b0ea-b076-4b97-b0f9-c4722343aac7",4826116]