From b4d25bf0cb2fa75733ae052c26d468ac5475aa8c Mon Sep 17 00:00:00 2001 From: Glyn Normington Date: Tue, 6 Aug 2024 17:20:26 +0100 Subject: [PATCH 1/3] Add FAQ --- FAQ.md | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) diff --git a/FAQ.md b/FAQ.md index 59a44db..b0d2167 100644 --- a/FAQ.md +++ b/FAQ.md @@ -2,7 +2,13 @@ ## How do we know AI companies/bots respect `robots.txt`? -The short answer is that we don't. `robots.txt` is a well-established standard but compliance is voluntary. There is no enforcement mechanism. +The short answer is that we don't. `robots.txt` is a well-established standard, but compliance is voluntary. There is no enforcement mechanism. + +## Why might AI web crawlers respect `robots.txt`? + +Larger and/or reputable companies developing AI models probably wouldn't want to damage their reputation by ignoring `robots.txt`. + +Also, given the contentious nature of AI and the possibility of legislation limiting its development, companies developing AI models will probably want to be seen to be behaving ethically, and so should (eventually) respect `robots.txt`. ## Can we block crawlers based on user agent strings? From 83d9397f176c781f388bac1765c09ff7a95c9477 Mon Sep 17 00:00:00 2001 From: "ai.robots.txt" Date: Tue, 6 Aug 2024 16:21:00 +0000 Subject: [PATCH 2/3] Removing previously generated files --- robots.txt | 32 -------------------------------- table-of-bot-metrics.md | 33 --------------------------------- 2 files changed, 65 deletions(-) delete mode 100644 robots.txt delete mode 100644 table-of-bot-metrics.md diff --git a/robots.txt b/robots.txt deleted file mode 100644 index 7f3cb46..0000000 --- a/robots.txt +++ /dev/null @@ -1,32 +0,0 @@ -User-agent: Amazonbot -User-agent: anthropic-ai -User-agent: Applebot-Extended -User-agent: Bytespider -User-agent: CCBot -User-agent: ChatGPT-User -User-agent: ClaudeBot -User-agent: Claude-Web -User-agent: cohere-ai -User-agent: Diffbot -User-agent: FacebookBot -User-agent: facebookexternalhit -User-agent: FriendlyCrawler -User-agent: Google-Extended -User-agent: GoogleOther -User-agent: GoogleOther-Image -User-agent: GoogleOther-Video -User-agent: GPTBot -User-agent: ICC-Crawler -User-agent: ImagesiftBot -User-agent: img2dataset -User-agent: Meta-ExternalAgent -User-agent: OAI-SearchBot -User-agent: omgili -User-agent: omgilibot -User-agent: PerplexityBot -User-agent: PetalBot -User-agent: Scrapy -User-agent: Timpibot -User-agent: VelenPublicWebCrawler -User-agent: YouBot -Disallow: / \ No newline at end of file diff --git a/table-of-bot-metrics.md b/table-of-bot-metrics.md deleted file mode 100644 index 2dafd43..0000000 --- a/table-of-bot-metrics.md +++ /dev/null @@ -1,33 +0,0 @@ -| Name | Operator | Respects `robots.txt` | Data use | Visit regularity | Description | -|-----|----------|-----------------------|----------|------------------|-------------| -| Amazonbot | Amazon | Yes | Service improvement and enabling answers for Alexa users. | No information. provided. | Includes references to crawled website when surfacing answers via Alexa; does not clearly outline other uses. | -| anthropic-ai | [Anthropic](https://www.anthropic.com) | Unclear at this time. | Scrapes data to train Anthropic's AI products. | No information. provided. | Scrapes data to train LLMs and AI products offered by Anthropic. | -| Applebot-Extended | [Apple](https://support.apple.com/en-us/119829#datausage) | Yes | Powers features in Siri, Spotlight, Safari, Apple Intelligence, and others. | Unclear at this time. | Apple has a secondary user agent, Applebot-Extended ... [that is] used to train Apple's foundation models powering generative AI features across Apple products, including Apple Intelligence, Services, and Developer Tools. | -| Bytespider | ByteDance | No | LLM training. | Unclear at this time. | Downloads data to train LLMS, including ChatGPT competitors. | -| CCBot | [Common Crawl](https://commoncrawl.org) | [Yes](https://commoncrawl.org/ccbot) | Provides crawl data for an open source repository that has been used to train LLMs. | Unclear at this time. | Sources data that is made openly available and is used to train AI models. | -| ChatGPT-User | [OpenAI](https://openai.com) | Yes | Takes action based on user prompts. | Only when prompted by a user. | Used by plugins in ChatGPT to answer queries based on user input. | -| ClaudeBot | [Anthropic](https://www.anthropic.com) | Unclear at this time. | Scrapes data to train Anthropic's AI products. | No information. provided. | Scrapes data to train LLMs and AI products offered by Anthropic. | -| Claude-Web | [Anthropic](https://www.anthropic.com) | Unclear at this time. | Scrapes data to train Anthropic's AI products. | No information. provided. | Scrapes data to train LLMs and AI products offered by Anthropic. | -| cohere-ai | [Cohere](https://cohere.com) | Unclear at this time. | Retrieves data to provide responses to user-initiated prompts. | Takes action based on user prompts. | Retrieves data based on user prompts. | -| Diffbot | [Diffbot](https://www.diffbot.com/) | At the discretion of Diffbot users. | Aggregates structured web data for monitoring and AI model training. | Unclear at this time. | Diffbot is an application used to parse web pages into structured data; this data is used for monitoring or AI model training. | -| FacebookBot | Meta/Facebook | [Yes](https://developers.facebook.com/docs/sharing/bot/) | Training language models | Up to 1 page per second | Officially used for training Meta "speech recognition technology," unknown if used to train Meta AI specifically. | -| facebookexternalhit | Meta/Facebook | [Yes](https://developers.facebook.com/docs/sharing/bot/) | No information. | Unclear at this time. | Unclear at this time. | -| FriendlyCrawler | Unknown | [Yes](https://imho.alex-kunz.com/2024/01/25/an-update-on-friendly-crawler) | We are using the data from the crawler to build datasets for machine learning experiments. | Unclear at this time. | Unclear who the operator is; but data is used for training/machine learning. | -| Google-Extended | Google | [Yes](https://developers.google.com/search/docs/crawling-indexing/overview-google-crawlers) | LLM training. | No information. | Used to train Gemini and Vertex AI generative APIs. Does not impact a site's inclusion or ranking in Google Search. | -| GoogleOther | Google | [Yes](https://developers.google.com/search/docs/crawling-indexing/overview-google-crawlers) | Scrapes data. | No information. | "Used by various product teams for fetching publicly accessible content from sites. For example, it may be used for one-off crawls for internal research and development." | -| GoogleOther-Image | Google | [Yes](https://developers.google.com/search/docs/crawling-indexing/overview-google-crawlers) | Scrapes data. | No information. | "Used by various product teams for fetching publicly accessible content from sites. For example, it may be used for one-off crawls for internal research and development." | -| GoogleOther-Video | Google | [Yes](https://developers.google.com/search/docs/crawling-indexing/overview-google-crawlers) | Scrapes data. | No information. | "Used by various product teams for fetching publicly accessible content from sites. For example, it may be used for one-off crawls for internal research and development." | -| GPTBot | [OpenAI](https://openai.com) | Yes | Scrapes data to train OpenAI's products. | No information. | Data is used to train current and future models, removed paywalled data, PII and data that violates the company's policies. | -| ICC-Crawler | [NICT](https://nict.go.jp) | Yes | Scrapes data to train and support AI technologies. | No information. | Use the collected data for artificial intelligence technologies; provide data to third parties, including commercial companies; those companies can use the data for their own business. | -| ImagesiftBot | [ImageSift](https://imagesift.com) | [Yes](https://imagesift.com/about) | ImageSiftBot is a web crawler that scrapes the internet for publicly available images to support our suite of web intelligence products | No information. | Once images and text are downloaded from a webpage, ImageSift analyzes this data from the page and stores the information in an index. Our web intelligence products use this index to enable search and retrieval of similar images. | -| img2dataset | [img2dataset](https://github.com/rom1504/img2dataset) | Unclear at this time. | Scrapes images for use in LLMs. | At the discretion of img2dataset users. | Downloads large sets of images into datasets for LLM training or other purposes. | -| Meta-ExternalAgent | [Meta](https://developers.facebook.com/docs/sharing/webmasters/web-crawlers) | Yes. | Used to train models and improve products. | No information. | "The Meta-ExternalAgent crawler crawls the web for use cases such as training AI models or improving products by indexing content directly." | -| OAI-SearchBot | [OpenAI](https://openai.com) | [Yes](https://platform.openai.com/docs/bots) | Search result generation. | No information. | Crawls sites to surface as results in SearchGPT. | -| omgili | [Webz.io](https://webz.io/) | [Yes](https://webz.io/blog/web-data/what-is-the-omgili-bot-and-why-is-it-crawling-your-website/) | Data is sold. | No information. | Crawls sites for APIs used by Hootsuite, Sprinklr, NetBase, and other companies. Data also sold for research purposes or LLM training. | -| omgilibot | [Webz.io](https://webz.io/) | [Yes](https://web.archive.org/web/20170704003301/http://omgili.com/Crawler.html) | Data is sold. | No information. | Legacy user agent initially used for Omgili search engine. Unknown if still used, `omgili` agent still used by Webz.io. | -| PerplexityBot | [Perplexity](https://www.perplexity.ai/) | [No](https://www.macstories.net/stories/wired-confirms-perplexity-is-bypassing-efforts-by-websites-to-block-its-web-crawler/) | Used to answer queries at the request of users. | Takes action based on user prompts. | Operated by Perplexity to obtain results in response to user queries. | -| PetalBot | [Huawei](https://huawei.com/) | Yes | Used to provide recommendations in Hauwei assistant and AI search services. | No explicit frequency provided. | Operated by Huawei to provide search and AI assistant services. | -| Scrapy | [Zyte](https://www.zyte.com) | Unclear at this time. | Scrapes data a variety of uses including training AI. | No information. | "AI and machine learning applications often need large amounts of quality data, and web data extraction is a fast, efficient way to build structured data sets." | -| Timpibot | [Timpi](https://timpi.io) | Unclear at this time. | Scrapes data for use in training LLMs. | No information. | Makes data available for training AI models. | -| VelenPublicWebCrawler | [Velen Crawler](https://velen.io) | [Yes](https://velen.io) | Scrapes data for business data sets and machine learning models. | No information. | "Our goal with this crawler is to build business datasets and machine learning models to better understand the web." | -| YouBot | [You](https://about.you.com/youchat/) | [Yes](https://about.you.com/youbot/) | Scrapes data for search engine and LLMs. | No information. | Retrieves data used for You.com web search engine and LLMs. | From 6d2285f5e0b391fc3646889dcb5a57989ba71623 Mon Sep 17 00:00:00 2001 From: "ai.robots.txt" Date: Tue, 6 Aug 2024 16:21:01 +0000 Subject: [PATCH 3/3] Add FAQ --- robots.txt | 32 ++++++++++++++++++++++++++++++++ table-of-bot-metrics.md | 33 +++++++++++++++++++++++++++++++++ 2 files changed, 65 insertions(+) create mode 100644 robots.txt create mode 100644 table-of-bot-metrics.md diff --git a/robots.txt b/robots.txt new file mode 100644 index 0000000..7f3cb46 --- /dev/null +++ b/robots.txt @@ -0,0 +1,32 @@ +User-agent: Amazonbot +User-agent: anthropic-ai +User-agent: Applebot-Extended +User-agent: Bytespider +User-agent: CCBot +User-agent: ChatGPT-User +User-agent: ClaudeBot +User-agent: Claude-Web +User-agent: cohere-ai +User-agent: Diffbot +User-agent: FacebookBot +User-agent: facebookexternalhit +User-agent: FriendlyCrawler +User-agent: Google-Extended +User-agent: GoogleOther +User-agent: GoogleOther-Image +User-agent: GoogleOther-Video +User-agent: GPTBot +User-agent: ICC-Crawler +User-agent: ImagesiftBot +User-agent: img2dataset +User-agent: Meta-ExternalAgent +User-agent: OAI-SearchBot +User-agent: omgili +User-agent: omgilibot +User-agent: PerplexityBot +User-agent: PetalBot +User-agent: Scrapy +User-agent: Timpibot +User-agent: VelenPublicWebCrawler +User-agent: YouBot +Disallow: / \ No newline at end of file diff --git a/table-of-bot-metrics.md b/table-of-bot-metrics.md new file mode 100644 index 0000000..2dafd43 --- /dev/null +++ b/table-of-bot-metrics.md @@ -0,0 +1,33 @@ +| Name | Operator | Respects `robots.txt` | Data use | Visit regularity | Description | +|-----|----------|-----------------------|----------|------------------|-------------| +| Amazonbot | Amazon | Yes | Service improvement and enabling answers for Alexa users. | No information. provided. | Includes references to crawled website when surfacing answers via Alexa; does not clearly outline other uses. | +| anthropic-ai | [Anthropic](https://www.anthropic.com) | Unclear at this time. | Scrapes data to train Anthropic's AI products. | No information. provided. | Scrapes data to train LLMs and AI products offered by Anthropic. | +| Applebot-Extended | [Apple](https://support.apple.com/en-us/119829#datausage) | Yes | Powers features in Siri, Spotlight, Safari, Apple Intelligence, and others. | Unclear at this time. | Apple has a secondary user agent, Applebot-Extended ... [that is] used to train Apple's foundation models powering generative AI features across Apple products, including Apple Intelligence, Services, and Developer Tools. | +| Bytespider | ByteDance | No | LLM training. | Unclear at this time. | Downloads data to train LLMS, including ChatGPT competitors. | +| CCBot | [Common Crawl](https://commoncrawl.org) | [Yes](https://commoncrawl.org/ccbot) | Provides crawl data for an open source repository that has been used to train LLMs. | Unclear at this time. | Sources data that is made openly available and is used to train AI models. | +| ChatGPT-User | [OpenAI](https://openai.com) | Yes | Takes action based on user prompts. | Only when prompted by a user. | Used by plugins in ChatGPT to answer queries based on user input. | +| ClaudeBot | [Anthropic](https://www.anthropic.com) | Unclear at this time. | Scrapes data to train Anthropic's AI products. | No information. provided. | Scrapes data to train LLMs and AI products offered by Anthropic. | +| Claude-Web | [Anthropic](https://www.anthropic.com) | Unclear at this time. | Scrapes data to train Anthropic's AI products. | No information. provided. | Scrapes data to train LLMs and AI products offered by Anthropic. | +| cohere-ai | [Cohere](https://cohere.com) | Unclear at this time. | Retrieves data to provide responses to user-initiated prompts. | Takes action based on user prompts. | Retrieves data based on user prompts. | +| Diffbot | [Diffbot](https://www.diffbot.com/) | At the discretion of Diffbot users. | Aggregates structured web data for monitoring and AI model training. | Unclear at this time. | Diffbot is an application used to parse web pages into structured data; this data is used for monitoring or AI model training. | +| FacebookBot | Meta/Facebook | [Yes](https://developers.facebook.com/docs/sharing/bot/) | Training language models | Up to 1 page per second | Officially used for training Meta "speech recognition technology," unknown if used to train Meta AI specifically. | +| facebookexternalhit | Meta/Facebook | [Yes](https://developers.facebook.com/docs/sharing/bot/) | No information. | Unclear at this time. | Unclear at this time. | +| FriendlyCrawler | Unknown | [Yes](https://imho.alex-kunz.com/2024/01/25/an-update-on-friendly-crawler) | We are using the data from the crawler to build datasets for machine learning experiments. | Unclear at this time. | Unclear who the operator is; but data is used for training/machine learning. | +| Google-Extended | Google | [Yes](https://developers.google.com/search/docs/crawling-indexing/overview-google-crawlers) | LLM training. | No information. | Used to train Gemini and Vertex AI generative APIs. Does not impact a site's inclusion or ranking in Google Search. | +| GoogleOther | Google | [Yes](https://developers.google.com/search/docs/crawling-indexing/overview-google-crawlers) | Scrapes data. | No information. | "Used by various product teams for fetching publicly accessible content from sites. For example, it may be used for one-off crawls for internal research and development." | +| GoogleOther-Image | Google | [Yes](https://developers.google.com/search/docs/crawling-indexing/overview-google-crawlers) | Scrapes data. | No information. | "Used by various product teams for fetching publicly accessible content from sites. For example, it may be used for one-off crawls for internal research and development." | +| GoogleOther-Video | Google | [Yes](https://developers.google.com/search/docs/crawling-indexing/overview-google-crawlers) | Scrapes data. | No information. | "Used by various product teams for fetching publicly accessible content from sites. For example, it may be used for one-off crawls for internal research and development." | +| GPTBot | [OpenAI](https://openai.com) | Yes | Scrapes data to train OpenAI's products. | No information. | Data is used to train current and future models, removed paywalled data, PII and data that violates the company's policies. | +| ICC-Crawler | [NICT](https://nict.go.jp) | Yes | Scrapes data to train and support AI technologies. | No information. | Use the collected data for artificial intelligence technologies; provide data to third parties, including commercial companies; those companies can use the data for their own business. | +| ImagesiftBot | [ImageSift](https://imagesift.com) | [Yes](https://imagesift.com/about) | ImageSiftBot is a web crawler that scrapes the internet for publicly available images to support our suite of web intelligence products | No information. | Once images and text are downloaded from a webpage, ImageSift analyzes this data from the page and stores the information in an index. Our web intelligence products use this index to enable search and retrieval of similar images. | +| img2dataset | [img2dataset](https://github.com/rom1504/img2dataset) | Unclear at this time. | Scrapes images for use in LLMs. | At the discretion of img2dataset users. | Downloads large sets of images into datasets for LLM training or other purposes. | +| Meta-ExternalAgent | [Meta](https://developers.facebook.com/docs/sharing/webmasters/web-crawlers) | Yes. | Used to train models and improve products. | No information. | "The Meta-ExternalAgent crawler crawls the web for use cases such as training AI models or improving products by indexing content directly." | +| OAI-SearchBot | [OpenAI](https://openai.com) | [Yes](https://platform.openai.com/docs/bots) | Search result generation. | No information. | Crawls sites to surface as results in SearchGPT. | +| omgili | [Webz.io](https://webz.io/) | [Yes](https://webz.io/blog/web-data/what-is-the-omgili-bot-and-why-is-it-crawling-your-website/) | Data is sold. | No information. | Crawls sites for APIs used by Hootsuite, Sprinklr, NetBase, and other companies. Data also sold for research purposes or LLM training. | +| omgilibot | [Webz.io](https://webz.io/) | [Yes](https://web.archive.org/web/20170704003301/http://omgili.com/Crawler.html) | Data is sold. | No information. | Legacy user agent initially used for Omgili search engine. Unknown if still used, `omgili` agent still used by Webz.io. | +| PerplexityBot | [Perplexity](https://www.perplexity.ai/) | [No](https://www.macstories.net/stories/wired-confirms-perplexity-is-bypassing-efforts-by-websites-to-block-its-web-crawler/) | Used to answer queries at the request of users. | Takes action based on user prompts. | Operated by Perplexity to obtain results in response to user queries. | +| PetalBot | [Huawei](https://huawei.com/) | Yes | Used to provide recommendations in Hauwei assistant and AI search services. | No explicit frequency provided. | Operated by Huawei to provide search and AI assistant services. | +| Scrapy | [Zyte](https://www.zyte.com) | Unclear at this time. | Scrapes data a variety of uses including training AI. | No information. | "AI and machine learning applications often need large amounts of quality data, and web data extraction is a fast, efficient way to build structured data sets." | +| Timpibot | [Timpi](https://timpi.io) | Unclear at this time. | Scrapes data for use in training LLMs. | No information. | Makes data available for training AI models. | +| VelenPublicWebCrawler | [Velen Crawler](https://velen.io) | [Yes](https://velen.io) | Scrapes data for business data sets and machine learning models. | No information. | "Our goal with this crawler is to build business datasets and machine learning models to better understand the web." | +| YouBot | [You](https://about.you.com/youchat/) | [Yes](https://about.you.com/youbot/) | Scrapes data for search engine and LLMs. | No information. | Retrieves data used for You.com web search engine and LLMs. |