AI大厂终于意识到继续与内容平台对抗是条不归路。 就在全球最知名的百科全书网站维基百科（Wikipedia）庆祝25周年之际，负责运营维基百科的维基媒 体基金会方面宣布，亚马逊、Meta、微软、Mistral AI以及Perplexity等多家AI大厂加入"维基媒体企业合 作伙伴计划"（Wikimedia Enterprise）。 因此也就意味着，这些厂商将付费获取维基百科的"企业级数据访问权"，以获取这家百科全书网站的实 时数据。而维基媒体企业合作伙伴计划则会根据他们的特定需求，对维基百科海量的文章数据进行结构 化处理，使其更易于模型训练和商业用途。对此维基媒体基金会表示，来自亚马逊、微软等厂商的授权 费用将直接用于支持该非营利组织的长期运营。 简而言之，维基百科将旗下的数据资产整理成AI更易懂的形式，以方便AI厂商即拿即用。 比如在金融大模型中，交易金额、交易时间、交易类型等结构化的交易记录，就可以作为模型的输入特 征，帮助AI学习和识别风险模式，从而提升输出结果的稳定性。不仅如此，结构化数据与知识图谱之 间存在天然的协同关系，通过将两者结合，AI大模型就可以更准确地理解数据的上下文和语义。 而维基百科之 ...

Core Insights - The Wikimedia Foundation announced a partnership with major tech companies including Amazon, Microsoft, and others to establish paid agreements for the use of Wikipedia content in AI model training, marking a shift from web scraping to legal data acquisition [1][4] Group 1: Partnership Details - The collaborations have been formalized over the past year and were previously kept confidential [4] - Companies will utilize Wikimedia's enterprise services to integrate human-governed knowledge resources on a large scale [4] Group 2: Financial Implications - The Wikimedia Foundation, as the largest non-profit encyclopedia platform, hosts over 65 million entries in more than 300 languages, making it a vital source of high-quality data for AI training [4] - The previous free scraping practices led to increased server load and operational costs, while the foundation's primary revenue relied on small public donations [4] - The new paid collaboration model provides a sustainable revenue stream for the non-profit organization [4] Group 3: Value of Knowledge - A spokesperson from the foundation emphasized that the long-term development of AI companies depends on the accumulation of human knowledge, and that paid partnerships respect knowledge producers [4]

Core Insights - Wikipedia has announced partnerships with major tech companies including Microsoft, Meta, and Amazon, marking a significant step in monetizing its content reliance by these firms [1][4] - The Wikimedia Foundation has signed agreements with several companies, including AI startups Perplexity and Mistral AI, in addition to existing partnerships [1][4] Industry Context - Wikipedia's content is crucial for training AI models, encompassing over 65 million entries in more than 300 languages, serving as a primary data source for tech giants developing generative AI chatbots and smart assistants [2][5] - The increasing demand for Wikipedia's free content for AI training has led to rising server demands and costs for the non-profit organization, which primarily relies on small public donations for funding [2][5] Business Model Evolution - The Wikimedia Foundation is promoting its enterprise service, which allows tech companies to pay for content training access and offers customized data services based on large-scale training needs [2][5] - Ryan Becker, president of Wikimedia Enterprise, emphasized the necessity for tech companies to financially support Wikipedia, recognizing the importance of transitioning from free access to commercial partnerships [6] Leadership Changes - The Wikimedia Foundation has appointed Bernadette Meehan, former U.S. ambassador to Chile, as the new CEO, effective January 20 [3][6]

来源：环球网 威尔士表示，他只是"略微"看过Grokipedia。当被问及对此有何看法时，威尔士提到了马斯克基于人工 智能的百科全书所引发的一些强烈反对意见。"我认为它所受到的很多批评对我来说并不意外。从信任 的角度来看，我不确定是否有人会信任一部带有偏见的百科全书。"他补充道。 威尔士试图将自己与马斯克区分开来。他说，如果他对维基百科上的某些内容不满意，他会发起讨论。 而马斯克则不同，威尔士暗示，马斯克很可能会自己去修改他不认同的内容。 外媒称，马斯克一贯喜欢插手公司细节，并自上而下地做出领导决策。 外媒还表示，维基百科无疑希望阻止用户转向像Grokipedia这样的竞争对手，因此允许人工智能机器人 抓取其内容。用户发现， Grokipedia上的一些页面几乎与维基百科的条目完全相同。（思瀚） 威尔士指的是埃隆·马斯克的 Grokipedia，这位特斯拉首席执行官在多年来指责维基百科存在偏见之后推 出了 Grokipedia。 【环球网科技综合报道】12月20日消息，据《商业内幕》报道，维基百科联合创始人吉米·威尔士表 示，人们可能不会信任"带有偏见"的百科全书。 ...

Core Viewpoint - The rise of AI-generated content has led to a significant decline in content quality, with the term "Slop" being coined to describe low-quality digital content produced en masse by AI [5][28]. Group 1: AI Content Proliferation - AI has transitioned from making basic errors to producing "correct but mediocre" content that lacks depth and engagement [5][12]. - The primary purpose of this AI-generated content is to deceive algorithms, fill ad spaces, and capture fleeting attention, rather than to provide meaningful information [7][26]. Group 2: Industry Response - Disney has entered the AI content space by partnering with OpenAI, investing $1 billion, and planning to integrate AI-generated videos into its Disney+ platform [10][12]. - The shift in the streaming industry has moved from a focus on high-quality content to a competition for content quantity, prompting major players like Disney to adopt AI for cost-effective content generation [12][14]. Group 3: Content Quality Divide - The internet is now divided into two realms: one that maintains a commitment to authenticity, such as Wikipedia and Spotify, and another that embraces "Slop" economics, prioritizing engagement metrics over quality [20][23]. - Platforms that encourage AI content production benefit from its low cost and speed, leading to a saturation of low-quality content that users must sift through to find valuable information [26][28].

Core Insights - The article discusses the transformation of knowledge in the internet age, highlighting both its unprecedented proliferation and potential crises [1][11] - It emphasizes the democratization of knowledge, where access and dissemination have become easier for the general public, moving away from elite control [4][12] - The article also addresses the challenges of information overload and the shifting authority of knowledge, necessitating new filtering mechanisms [5][6][10] Group 1: Characteristics of Networked Knowledge - Networked knowledge is characterized by richness, with vast amounts of information available online, such as over 40 billion indexed web pages and 7 million Wikipedia entries [2][3] - The internet has enabled a significant increase in the quantity of accessible knowledge, with projections of around 1.2 billion websites by January 2025 [2] - Knowledge is no longer confined to traditional institutions but is now a dynamic entity that evolves with real-time contributions from users [4][18] Group 2: Knowledge Democratization - The article outlines the historical context of knowledge democratization, tracing back to the invention of the printing press and the role of public libraries [4] - Digital technology has further propelled this democratization, allowing knowledge to transcend geographical and social barriers [4][12] - Wikipedia serves as a prime example of this phenomenon, providing an open platform for knowledge sharing and real-time updates [4][2] Group 3: Knowledge Filtering Mechanisms - The overwhelming abundance of information has led to a need for effective filtering systems to manage knowledge [5][6] - Traditional filtering methods have lost some authority, and new mechanisms, including algorithmic and social filtering, are being explored [7][10] - The article warns of potential issues with these new filtering systems, such as echo chambers and confirmation bias, which can distort knowledge acquisition [10][9] Group 4: Future Directions for Knowledge - The article advocates for open access initiatives to remove barriers to knowledge, allowing free and unrestricted access to academic research [12][13] - It emphasizes the importance of linking knowledge across platforms and fostering a culture of sharing and collaboration [15][16] - Education on critical thinking and effective internet use is deemed essential for navigating the complexities of networked knowledge [19][20]

Core Insights - PolitiFact's evaluation of xAI's Grokipedia reveals serious issues including content plagiarism and lack of factual accuracy, damaging its credibility as a reliable information source [2][3][4] - Grokipedia, launched in October 2023, aimed to compete with Wikipedia but has faced widespread criticism for its AI-generated content that often replicates Wikipedia without proper attribution [2][3] Content Plagiarism - PolitiFact analyzed 885,279 articles on Grokipedia, finding that many directly copied text from Wikipedia, with a similarity rate as high as 96% in some cases [2][4] - An example includes the "Monday" entry, which omitted 22 reference sources from Wikipedia while claiming to adapt content under a Creative Commons license [4] Lack of Factual Accuracy - Grokipedia's AI-generated content has been found to introduce errors, such as a misleading statement regarding the Nobel Prize, which lacked citation and was factually incorrect [4][6] - The platform often simplifies narratives by removing context from Wikipedia, sacrificing accuracy in the process [4][5] Transparency Issues - Grokipedia's editing mechanism lacks transparency, with no visible edit history or clear error correction processes, making its "fact-checking" label difficult to trust [6] - In contrast, Wikipedia maintains a transparent editing process, allowing anyone to edit and providing a complete history of article changes [5][6] Industry Context - The launch of Grokipedia coincides with xAI's legal disputes with Apple and OpenAI, which have accused these companies of suppressing AI competition [6] - The Wikimedia Foundation has raised concerns about Grokipedia's plagiarism and the implications for educational ecosystems, emphasizing the need for responsible use of open-source content [6]

Core Points - Elon Musk has criticized Wikipedia multiple times and has launched his own AI-driven online encyclopedia called GrokiPedia, which aims to surpass Wikipedia [1][4] - Wikipedia has responded to Musk's criticisms by emphasizing its human-operated nature and non-profit status, contrasting it with Musk's wealth and the commercial nature of large tech companies [1][4] Group 1 - Wikipedia has been operational for nearly 25 years and is committed to being a human-created platform rather than a machine-generated one [4] - Wikipedia's content is maintained by volunteers, while GrokiPedia is developed by xAI and utilizes the Grok large language model for generating and editing entries [4] - GrokiPedia 0.1 was officially launched recently, claiming to have access to 885,279 articles, with Musk stating that it is a necessary step for achieving xAI's goal of understanding the universe [4]

Core Points - Elon Musk's xAI company launched Grokipedia, an AI-generated competitor to Wikipedia, with a promise that the upcoming version 1.0 will be "ten times better" than the current version [1] - Grokipedia has over 885,000 articles as of October 29, 2023, but its content is primarily AI-generated, contrasting with Wikipedia's nearly 8 million articles created by human contributors [1] - Musk has criticized Wikipedia for perceived biases and has called for a cessation of funding to the platform [1] Group 1 - Grokipedia's interface is simple and similar to search engines like Google [1] - The platform is positioned as a source for "truth," with Musk acknowledging the challenge of achieving perfection [1] - Critics argue that Grokipedia's content often mirrors Wikipedia's and is heavily reliant on unverified AI-generated material, suggesting it reflects Musk's viewpoints rather than objective facts [2] Group 2 - The Wikimedia Foundation emphasized that Wikipedia will continue to be human-operated and highlighted that human-created knowledge is foundational for AI-generated content [2] - Scholars suggest that Musk's launch of Grokipedia is a strategic move to gain control over narrative and discourse [2]

Core Viewpoint - Elon Musk's AI company xAI has launched an AI-driven encyclopedia website to compete with Wikipedia, featuring over 885,000 articles compared to Wikipedia's over 7 million articles [1][3]. Group 1: Product Features - The homepage of Musk's encyclopedia is very simple, displaying only the title and a search box for user queries [3]. - The website is currently in version 0.1, with plans for an upgrade to version 1.0 that will enhance its functionality [3]. Group 2: Market Context - Musk has criticized Wikipedia for containing biased information, although some entries on his encyclopedia cite Wikipedia as a source [3]. - The Wikimedia Foundation, which operates Wikipedia, has stated that previous alternatives to Wikipedia have not disrupted its operations [3].