Google 搜尋的 AI 產生內容相關指引
透過集合功能整理內容
你可以依據偏好儲存及分類內容。
2023 年 2 月 8 日，星期三
長久以來，Google 一直深信 AI 具有足夠潛力，能夠改變系統提供實用資訊的服務能力。本文將進一步介紹 AI 產生的內容如何搭配我們長期採用的方法，為 Google 搜尋使用者呈現實用內容。
獎勵優質內容，無論其製作方式為何
Google 的排名系統旨在獎勵原創且高品質的內容，這類內容需達到我們所謂的 E-E-A-T 標準：專業性、實際經驗、權威性和可信度。歡迎前往 Google 搜尋的運作方式網站瞭解詳情。
我們將重點放在內容品質而非內容製作方式，多年來，這項原則有助於我們為使用者提供可靠的高品質搜尋結果。
舉例來說，大約在 10 年前，由人類生產的大量製作內容開始興起，人們產生了合理的擔憂。但沒有人認為，我們應該宣告禁止所有由人類生產的內容來回應這種趨勢。轉而改善系統，獎勵優質內容顯然是更合理的做法，這也正是我們在做的。
從 Google 起步以來，重視獎勵優質內容一直是我們的核心原則。此一原則延續至今，包括透過用來凸顯可靠資訊的排名系統，以及我們的實用內容系統，都屬於這類做法。我們去年推出的實用內容系統，目的是確保搜尋時得到的結果，主要是為使用者建立的內容，而非為影響搜尋排名而建立的內容。
自動化功能如何製作實用內容
針對自動產生的內容，我們多年來一直遵循同樣的指引。使用自動化功能 (包括 AI) 產生內容時，如果主要目的是操控搜尋引擎中的排名，這是違反垃圾內容政策的行為。
多年來，Google 已經有相當豐富的經驗，對付採用自動化功能試圖影響遊戲搜尋結果的做法。無論垃圾內容是怎麼製造出來的，我們杜絕垃圾內容的措施 (包括 SpamBrain 系統) 都會持續運作，
話雖如此，請務必瞭解，並非所有使用自動化功能產生的內容 (包括 AI 產生內容) 都是垃圾內容。使用自動化功能產生實用內容的做法已經行之有年，例如體育賽事比分、天氣預報和轉錄稿。AI 能夠提供全新層次的表達方式與創意，並成為協助使用者製作優質網路內容的重要工具。
這完全符合 Google 一直以來的方針，希望讓使用者順利發揮新科技的力量。我們會繼續採取這種負責任的做法，同時針對資訊品質和 Google 搜尋內容的整體實用性，維持嚴格的標準。
針對有意運用 AI 生成技術的創作者，以下是我們的建議：
如前所述，無論內容製作途徑為何，任何人想在 Google 搜尋中獲得良好成效，都必須致力於製作原創性、高品質，且以使用者為優先的內容，也必須達到 E-E-A-T 標準。
如要進一步瞭解 E-E-A-T 的概念，創作者可參閱製作實用、可靠，且以使用者為優先的內容說明頁面。此外，我們也更新了該網頁，從對象、方法與原因的角度來思考內容製作，並提供相關指引。
無論您使用的是否為 AI 產生的內容，只要以這種方式評估內容，就能協助您把握 Google 系統所要獎勵的特點。
發文者：Google 搜尋品質團隊 Danny Sullivan 和 Chris Nelson
常見問題
為進一步協助您，以下列出一些有關 AI 內容和 Google 搜尋的常見問題，並提供答案：
AI 內容是否有違 Google 搜尋的規範？
適當使用 AI 或自動化功能並不會違反我們的規範。意思是不得使用這類功能產生主要為操控搜尋排名而製作的內容，這類內容違反我們的垃圾內容政策。
為什麼 Google 搜尋不禁止 AI 內容？
使用自動化功能建立實用內容的做法已經實行很久。AI 能以新穎的方式提供協助並產生實用內容。
Google 搜尋如何防止品質不佳的 AI 內容充斥於搜尋結果？
劣質內容對 Google 搜尋來說並非新挑戰。多年來，我們一直都在處理由人類和自動化功能建立的劣質內容我們目前已有一些系統可以判定內容實用程度。還有其他系統用來提升原創新聞報導的能見度。我們的系統會持續定期改善。
Google 會如何處理可能會針對重要議題散播不實資訊，或是牴觸普遍共識的 AI 內容？
這些問題在人類產生的內容和 AI 產生的內容中都存在。無論內容是如何製作出來，針對重要議題，我們的系統都會設法凸顯來自可靠來源的高品質資訊，而非違反可靠共識的資訊。對於主題品質極為重要的議題 (例如健康、公民或財務資訊)，我們的系統會更加重視信號的可靠性。
Google 搜尋如何判斷 AI 是否在搜尋結果中製造垃圾內容？
我們有各式各樣的系統 (包括 SpamBrain) 可以分析模式和信號，協助我們識別垃圾內容 (無論其製作途徑為何)。
AI 內容在 Google 搜尋中的排名是否會比較高？
我該使用 AI 來產生內容嗎？
如果對您來說，AI 有助於製作實用且原創的內容，那麼就值得考慮採用。如果您是把 AI 當成一種廉價又簡單的操控技術，可以影響搜尋引擎排名，則請勿使用 AI。
我是否應該為所有內容新增作者署名？
在讀者合理期望得知作者資訊的情況下 (例如任何會讓人想知道「誰寫了這篇文章？」的內容)，建議您提供正確的作者署名。
提醒您，顯示在 Google 新聞中的發布者應使用署名和作者資訊。詳情請參閱 Google 新聞政策頁面。
我應該在內容中加入 AI 或自動化功能揭露資訊嗎？
如果您的內容可能會讓人疑惑「這是怎麼製作出來的？」，加入 AI 或自動化功能揭露資訊就會很有實用性。理應讓人得知這類資訊的情況下，請加入相關資訊。
我可以將 AI 列為內容作者嗎？
我們的建議是讓讀者清楚知道，在產生此內容的程序中有 AI 參與，但在作者署名處填上 AI 可能不是最好的方式。
除非另有註明，否則本頁面中的內容是採用創用 CC 姓名標示 4.0 授權，程式碼範例則為阿帕契 2.0 授權。詳情請參閱《Google Developers 網站政策》。Java 是 Oracle 和/或其關聯企業的註冊商標。
[null,null,[],[[["\u003cp\u003eGoogle Search prioritizes high-quality content that demonstrates E-E-A-T (expertise, experience, authoritativeness, and trustworthiness), regardless of how it's created, including by AI.\u003c/p\u003e\n"],["\u003cp\u003eUsing AI to generate content solely for manipulating search rankings is against Google's spam policies.\u003c/p\u003e\n"],["\u003cp\u003eGoogle's systems are designed to identify and demote spam, including content generated through automation, while rewarding original, helpful content.\u003c/p\u003e\n"],["\u003cp\u003eContent creators should focus on producing original, high-quality, people-first content, whether using AI or not, to succeed in Google Search.\u003c/p\u003e\n"],["\u003cp\u003eTransparency about the use of AI in content creation is encouraged, with disclosures and appropriate bylines where relevant.\u003c/p\u003e\n"]]],["Google focuses on rewarding high-quality content, regardless of whether it's human or AI-generated. Using AI to manipulate search rankings violates Google's spam policies, but AI's use for creating helpful content is acceptable. Creators should prioritize original, people-first content demonstrating expertise, experience, authoritativeness, and trustworthiness (E-E-A-T). Google's systems evaluate content helpfulness and use signals to identify spam. AI-generated content should include clear disclosures and author information when reasonably expected.\n"],null,["# Google Search's guidance about AI-generated content\n\nWednesday, February 8, 2023\n\n\nAt Google, we've long believed in the power of AI to transform the ability to deliver helpful\ninformation. In this post, we'll share more about how AI-generated content fits into our\nlong-standing approach to show helpful content to people on Search.\n\nRewarding high-quality content, however it is produced\n------------------------------------------------------\n\n\nGoogle's ranking systems aim to reward original, high-quality content that demonstrates qualities\nof what we call E-E-A-T: expertise, experience, authoritativeness, and trustworthiness. We share\nmore about this [in our How Search Works site](https://www.google.com/search/howsearchworks/how-search-works/rigorous-testing/).\n\n\nOur focus on the quality of content, rather than how content is produced, is a useful guide that\nhas helped us deliver reliable, high quality results to users for years.\n\n\nFor example, about 10 years ago, there were understandable concerns about a rise in mass-produced\nyet human-generated content. No one would have thought it reasonable for us to declare a ban on\nall human-generated content in response. Instead, it made more sense to improve our systems to\nreward quality content, [as we did](https://googleblog.blogspot.com/2011/02/finding-more-high-quality-sites-in.html).\n\n\nFocusing on rewarding quality content has been core to Google since we began. It continues today,\nincluding through our [ranking systems](/search/docs/appearance/ranking-systems-guide)\ndesigned to [surface reliable information](/search/docs/appearance/ranking-systems-guide#reliable-information-systems)\nand our [helpful content system](/search/updates/helpful-content-update). The helpful\ncontent system was introduced last year to better ensure those searching get content created\nprimarily for people, rather than for search ranking purposes.\n\nHow automation can create helpful content\n-----------------------------------------\n\n\nWhen it comes to automatically generated content, our guidance has been consistent for years.\nUsing automation---including AI---to generate content with the primary purpose of manipulating\nranking in search results is a [violation of our spam policies](/search/docs/essentials/spam-policies#scaled-content).\n\n\nGoogle has many years of experience dealing with automation being used in an attempt to game search\nresults. Our spam-fighting efforts---including our [SpamBrain system](/search/blog/2022/04/webspam-report-2021)---will\ncontinue, however spam is produced.\n\n\nThis said, it's important to recognize that not all use of automation, including AI generation, is\nspam. Automation has long been used to generate helpful content, such as sports scores, weather\nforecasts, and transcripts. AI has the ability to power new levels of expression and creativity,\nand to serve as a critical tool to help people create great content for the web.\n\n\nThis is in line with how we've always thought about empowering people with new technologies. We'll\ncontinue taking this responsible approach, while also maintaining a high bar for information quality\nand the overall helpfulness of content on Search.\n\nOur advice for creators considering AI-generation\n-------------------------------------------------\n\n\nAs explained, however content is produced, those seeking success in Google Search should be looking\nto produce original, high-quality, people-first content demonstrating qualities E-E-A-T.\n\n\nCreators can learn more about the concept of E-E-A-T on our\n[Creating helpful, reliable, people-first content](/search/docs/fundamentals/creating-helpful-content)\nhelp page. In addition, we've updated that page with some guidance about thinking in terms of\n[Who, How, and Why](/search/docs/fundamentals/creating-helpful-content#ask-who-how-why)\nin relation to how content is produced.\n\n\nEvaluating your content in this way, whether you're using AI-generated content or not, will help\nyou stay on course with what our systems seek to reward.\n\nPosted by [Danny Sullivan](https://mastodon.social/@searchliaison)\nand [Chris Nelson](https://www.linkedin.com/in/nelso/),\non behalf of the Google Search Quality team\n\nFAQ\n---\n\n\nTo further help, here are some answers to questions you may have about AI content and Google Search: \n\n### Is AI content against Google Search's guidelines?\n\n\nAppropriate use of AI or automation is not against our guidelines. This means that it is not used\nto generate content primarily to manipulate search rankings, which is\n[against our spam policies](/search/docs/essentials/spam-policies#scaled-content). \n\n### Why doesn't Google Search ban AI content?\n\n\nAutomation has long been used in publishing to create useful content. AI can assist with and\ngenerate useful content in exciting new ways. \n\n### How will Google Search prevent poor quality AI content from taking over search results?\n\n\nPoor quality content isn't a new challenge for Google Search to deal with. We've been tackling poor\nquality content created both by humans and automation for years. We have existing\n[systems](/search/docs/appearance/ranking-systems-guide) to determine the\n[helpfulness of content](/search/updates/helpful-content-update). Other systems work\nto elevate [original news reporting](https://blog.google/products/search/original-reporting/).\nOur systems continue to be\n[regularly improved](https://status.search.google.com/products/rGHU1u87FJnkP6W2GwMi/history). \n\n### How will Google address AI content that potentially propagates misinformation or contradicts\nconsensus on important topics?\n\n\nThese issues exist in both human-generated and AI-generated content. However content is produced,\nour systems [look to surface high-quality information](/search/docs/appearance/ranking-systems-guide#reliable-information-systems)\nfrom reliable sources, and not information that contradicts well-established consensus on important\ntopics. On topics where information quality is critically important---like health, civic, or\nfinancial information---our systems place an even greater emphasis on signals of reliability. \n\n### How can Search determine if AI is being used to spam search results?\n\n\nWe have a variety of systems, including [SpamBrain](/search/blog/2022/04/webspam-report-2021),\nthat analyze patterns and signals to help us identify spam content, however it is produced. \n\n### Will AI content rank highly on Search?\n\n\nUsing AI doesn't give content any special gains. It's just content. If it is\n[useful, helpful, original, and satisfies aspects of E-E-A-T](/search/docs/fundamentals/creating-helpful-content),\nit might do well in Search. If it doesn't, it might not. \n\n### Should I use AI to generate content?\n\n\nIf you see AI as an essential way to help you produce content that is helpful and original, it\nmight be useful to consider. If you see AI as an inexpensive, easy way to game search engine\nrankings, then no. \n\n### Should I add author bylines to all my content?\n\n\nYou should consider having accurate author bylines when readers would reasonably expect it, such\nas to any content where someone might think, \"Who wrote this?\"\n\n\nAs a reminder, publishers that appear in Google News should use bylines and author information.\nLearn more on our [Google News policies](https://support.google.com/news/publisher-center/answer/6204050)\npage. \n\n### Should I add AI or automation disclosures to my content?\n\n\nAI or automation disclosures are useful for content where someone might think \"How was this created?\".\nConsider adding these when it would be reasonably expected. \n\n### Can I list AI as the author of content?\n\n\nGiving AI an author byline is probably not the best way to follow our recommendation to\n[make clear to readers](/search/docs/fundamentals/creating-helpful-content#how-the-content-was-created)\nwhen AI is part of the content creation process."]]
