ai_society Social Science

6 个科学问题 · 社会科学 Social Science

本页目录

算法治理有哪些主要路径?

What are the main paths of algorithmic governance?

路径包括事前审计与影响评估、透明度披露、行业自律,以及针对高风险应用的强制许可与人工监督,目标在创新与安全之间平衡。
Approaches include pre-deployment audits and impact assessments, transparency disclosure, self-regulation, and mandatory licensing with human oversight for high-risk uses, balancing innovation and safety.
难度 专业 · 适合 高中(15-18) CC BY 4.0 · 已核验 · Verified 2026-11-06
来源 · Source

欧盟委员会《通用数据保护条例》(GDPR, 2018)

European Commission, General Data Protection Regulation (GDPR, 2018)

#算法治理 #AI监管 #风险分级

自动化与AI对就业结构有何结构性影响?

What are the structural effects of automation and AI on employment?

AI更可能替代常规性任务而非整体岗位,导致“就业极化”——中端常规岗位收缩、高低技能岗位扩张,并加剧技能溢价与收入不平等。
AI tends to substitute routine tasks rather than whole jobs, producing “job polarization”: middle-skill routine roles shrink while high- and low-skill roles grow, widening skill premiums and inequality.
难度 专业 · 适合 高中(15-18) CC BY 4.0 · 已核验 · Verified 2026-08-23 · 2 条独立证据
来源 · Source

Manuel Castells,《The Rise of the Network Society》Blackwell,1996 初版;2000 第2版

M. Castells, The Rise of the Network Society, Blackwell, 1st ed. 1996, 2nd ed. 2000

#自动化 #就业极化 #技能溢价

AI普及会扩大还是缩小数字鸿沟?

Does AI widen or narrow the digital divide?

数字鸿沟指不同群体在接入和使用数字技术上的差距。AI既可降低使用门槛(如翻译、辅助),也可能因算力与技能门槛而扩大差距,取决于可及性政策。
The digital divide is unequal access to and use of digital technology. AI can lower barriers (translation, assistance) but may widen gaps via compute and skill requirements, depending on accessibility policy.
难度 进阶 · 适合 大众 CC BY 4.0 · 已核验 · Verified 2026-08-23 · 2 条独立证据
来源 · Source

Manuel Castells,《The Rise of the Network Society》Blackwell,1996 初版;2000 第2版

M. Castells, The Rise of the Network Society, Blackwell, 1st ed. 1996, 2nd ed. 2000

#数字鸿沟 #可及性 #技术不平等

什么是“深度伪造”(deepfake)?它对信息信任有何影响?

What are deepfakes, and how do they affect trust in information?

深度伪造利用生成式AI合成逼真的音视频,可被用于误导与操纵。它侵蚀“眼见为实”的信任基础,推动事实核查与内容溯源技术。
Deepfakes use generative AI to fabricate realistic audio/video for deception, eroding the “seeing is believing” trust basis and spurring fact-checking and content provenance.
难度 进阶 · 适合 小学高段(9-12) CC BY 4.0 · 来源待核验 · Source pending review
来源 · Source

出处备注:此条未指向带版次的可引用著作(通识性内容、机构文件或未展开的英文书名),本库不提供独立书目出处;如需精确引用,需由编辑补全具体篇目与版次。

中国科学院(网络谣言传播研究)

Chinese Academy of Sciences (research on online rumor propagation)

#深度伪造 #虚假信息 #生成式AI

什么是“黑箱”模型?可解释AI为何重要?

What are “black-box” models, and why does explainable AI matter?

复杂模型(如深度神经网络)的决策过程难以被人理解,称为黑箱。在医疗、司法等高风险场景,可解释AI帮助人们审查、追责并建立信任。
Complex models like deep neural networks make decisions hard to interpret (“black box”). In high-stakes domains like medicine and justice, explainable AI aids review, accountability, and trust.
难度 进阶 · 适合 初中(12-15) CC BY 4.0 · 已核验 · Verified 2026-11-04
来源 · Source

Rudin, Radin,《Why Are We Using Black Box Models in AI When We Don’t Need To? A Lesson From An Explainable AI Competition》,Harvard Data Science Review,2019 (DOI: 10.1162/99608f92.5a8a3a3d)

Rudin, Radin, 'Why Are We Using Black Box Models in AI When We Don’t Need To? A Lesson From An Explainable AI Competition', Harvard Data Science Review, 2019, DOI: 10.1162/99608f92.5a8a3a3d

#黑箱模型 #可解释AI #问责

机器人能“思考”吗?

Can robots “think”?

现在的AI能根据数据做很多事,比如认图片、回答问题,但它不会像人一样真正理解和感受,更像很会模仿的助手。
Today’s AI can do many tasks from data—recognize images, answer questions—but it doesn’t truly understand or feel like people do; it is more like a skilled imitator.
难度 入门 · 适合 幼儿(3-6) CC BY 4.0 · 来源待核验 · Source pending review
来源 · Source

出处备注:著录含书名,但缺版次 / 出版社 / 年份中的一部分,需编辑补全后才能做独立比对。

图灵 1950 年论文《计算机器与智能》

Turing's 1950 paper Computing Machinery and Intelligence

#人工智能 #机器人 #思考