What Apple says Siri AI does
Apple's announcement of 14 September describes Siri AI as an entirely new Siri with “broad world knowledge”. As well as acting on a user's own messages, email and photos, it “can also use broad world knowledge to get up-to-date information on virtually any topic”, and “will leverage information from across the web to deliver helpful, accurate answers”. Users can ask by voice, from the side button, or on iPad and Mac through Spotlight, which Apple says lets them “search for answers to almost any question”. A new Siri app keeps a history of conversations across devices.
For a company, the relevant use is the one Apple gives as an example: questions about the world, answered with information gathered from the web. Someone who asks Siri about a business, a product or an executive may now get a written answer rather than a list of results. The announcement says nothing about which websites those answers rely on, how they are ranked, or whether they link to their sources.
What Apple's crawler documentation says
Apple's support page “About Applebot” is more specific. Applebot is Apple's web crawler, and its data powers “the search technology integrated into many user experiences in Apple's ecosystem including Spotlight, Siri, and Safari”. The page then makes three separate statements.
Search. Allowing Applebot in robots.txt “allows website content to appear in search results for Apple users around the world” in those products. Blocking it removes a site from them.
Training. Applebot data “may also be used to help train Apple foundation models”. Publishers can opt out by disallowing a separate token, Applebot-Extended. Apple is explicit that this does not affect search: even with Applebot-Extended disallowed, a site's content “will remain discoverable through Spotlight, Siri, and Safari”.
AI answers. Applebot data “may be used to provide additional context and up-to-date content when AI models are used to generate output for display in Apple products and services”, for example “answering broad world knowledge questions in Siri and Search that may include links to sources and websites used to help generate the answer”. The opt-out for this is not in robots.txt. It is the nosnippet meta tag: with it, Applebot “won't generate a description or web answer for the page”, and Apple “will not use data tagged nosnippet as additional context” for AI output.
Two further details matter. If a robots.txt file does not mention Applebot but does mention Googlebot, Apple says Applebot follows the Googlebot rules. And a site that blocks Applebot-Extended and uses nosnippet can still be crawled by Applebot and appear in Spotlight, Siri and Safari, with a title but no description.
What 30 news sites allow
Because Apple splits training from answering, the question for a company is whether the outlets that cover it let Applebot in. On 5 October 2026 we read the robots.txt files of 30 large news and business publications, chosen as outlets companies commonly seek coverage in. All 30 returned a file.
Applebot-Extended (training). 22 of the 30 disallow it from the whole site: the New York Times, the Washington Post, the Guardian, the BBC, CNN, Reuters, AP, Bloomberg, Forbes, CNBC, the Financial Times, The Verge, TechCrunch, Wired, The Atlantic, Inc., Fast Company, USA Today, NPR, NBC News, The Economist and Newsweek. The other eight do not mention it.
Applebot (search and answers). Only USA Today blocks it, with User-agent: Applebot followed by Disallow: /. Four sites name Applebot and allow it: the Wall Street Journal, Reuters, The Verge and The Atlantic. Seven, including the New York Times, CNBC, Fortune and NPR, do not name it but set rules for Googlebot, which Applebot then follows; none of those rules block the home page. The remaining 18 leave Applebot to their general rules for all crawlers, which in every case allow the home page.
Google-Extended (Google's training control), for comparison. 15 of the 30 disallow it.
Number of the 30 sites whose robots.txt names the crawler and disallows it from the whole site, read October 5, 2026, rechecked October 10, 2026
Source: robots.txt files of the 30 domains listed in the methodology; original measurement.
The Wall Street Journal does not name either training token but blocks both through its catch-all rule for unlisted crawlers; counting that, Applebot-Extended is blocked on 23 sites. The October 10 recheck matched the October 5 counts.
Number of the 30 sites, by the rule Applebot follows under Apple's documentation, read October 5, 2026, rechecked October 10, 2026
Source: robots.txt files of the 30 domains listed in the methodology; Apple Support, About Applebot, for the Googlebot fallback.
Allow and block refer to the site's home page only.
So for 29 of the 30, Apple's crawler is permitted, and coverage there is available, on Apple's own account, to inform Siri's answers, while most of the same outlets refuse to let Apple train on it. That is the same split publishers make with OpenAI, where most block training crawlers but admit search ones, as set out in do news sites block AI crawlers. USA Today's coverage is the exception: by its robots.txt, it is closed to Siri and Spotlight altogether.
What it means for a company
Coverage counts in one more place. Placements in outlets that admit Applebot are, on Apple's description, material Siri can draw on when someone asks about a company. Apple has not said how it weighs them against a company's own site or other sources.
Check your own site. A company's own robots.txt should be read with Apple's fallback rule in mind. A site that has no Applebot rule but blocks Googlebot from a section will block Applebot from it too. A site that wants to stay out of Apple's training but in Siri should disallow Applebot-Extended, not Applebot. How blocking one crawler can take a site out of search is covered in does blocking a crawler remove you from search.
Nosnippet has a cost. The tag that keeps a page out of AI answers also removes its description from Apple's search results, and Google's documentation lists the same tag among the controls that limit what its own Search, including its AI features, shows from a page. It is a blunt tool for a company's own pages and one a company cannot apply to anyone else's.
No referral data yet. Apple has not documented a referrer or link tag for visits from Siri answers, so there is no published way to count them in analytics.
Methodology and limitations
Apple's statements are quoted from its newsroom announcement “Siri AI, a profoundly more capable and personal assistant powered by the next generation of Apple Intelligence, is here” of 14 September 2026 and its support page “About Applebot”, both read on 5 October 2026. The robots.txt reading is an original measurement. At about 15:50 UTC on 5 October 2026 we requested https://www.<domain>/robots.txt for 30 domains (nytimes.com, washingtonpost.com, wsj.com, theguardian.com, bbc.com, cnn.com, reuters.com, apnews.com, bloomberg.com, forbes.com, businessinsider.com, cnbc.com, ft.com, theverge.com, techcrunch.com, wired.com, theatlantic.com, axios.com, fortune.com, inc.com, fastcompany.com, entrepreneur.com, usatoday.com, latimes.com, npr.org, nbcnews.com, foxnews.com, time.com, economist.com and newsweek.com), following redirects, with a desktop Safari user agent. We parsed grouped user-agent blocks and judged whether each agent was disallowed from the site root, using the longest matching rule. For Applebot we used a group naming Applebot if there was one, otherwise a Googlebot group, as Apple's documentation describes, otherwise the general group. Our first pass misread Forbes, whose server sent a compressed file our request had not asked to be decompressed; we re-ran all 30 with decompression and these figures come from that run. Five limits apply. We judged access to the home page only, not to sections. Robots.txt records stated permission at one moment from one vantage point, and does not show what Applebot actually crawled. We did not check article pages for nosnippet tags. The 30 domains are a purposive sample of large English-language outlets, not a representative sample. And Apple's documentation describes what Applebot data “may” be used for; we have not tested Siri AI, which requires a supported device, and cannot say which sources it actually uses or how often it links to them.
Conclusion
Apple has given publishers and companies a clearer map than most AI providers: Applebot for search, Applebot-Extended for training, nosnippet for AI answers. Major news publishers have used it the way they use OpenAI's: 22 of the 30 we checked refuse Apple's training, but 29 still let Applebot in, so their coverage remains available for Siri's answers on Apple's account. For a company, that means placements in most large outlets can now inform one more AI assistant, built into Apple's devices, and that a careless robots.txt rule written for Google can shut its own site out of Siri. What Apple has not said is how Siri chooses among those sources, and until it does, the best that can be checked is whether a page is allowed in.
Frequently Asked Questions
Where does Siri AI get information about companies?
Apple says Siri AI uses information from across the web for broad world-knowledge questions, and its Applebot documentation says data collected by Applebot may be used as context for those answers, which may include links to sources. Apple has not said how it chooses among sources.
Does blocking Applebot-Extended keep my content out of Siri?
No. Apple says Applebot-Extended controls only whether content is used to train its foundation models. Content stays discoverable through Spotlight, Siri and Safari.
How do I keep a page out of Siri's AI answers?
Apple says to use the nosnippet meta tag. Applebot will then not generate a description or web answer for the page, and Apple will not use it as context for AI output. The page can still appear with its title.
Do news sites let Applebot in?
Of 30 major news and business sites we checked on 5 October 2026, 29 allowed Applebot to crawl their home page. Only USA Today blocked it. 22 blocked Applebot-Extended, Apple's training control.
What if my robots.txt does not mention Applebot?
Apple says Applebot then follows any rules you have written for Googlebot. If there are none, your general rules for all crawlers apply.
Can I see visits from Siri in my analytics?
Apple has not documented a referrer or tracking tag for visits from Siri answers, so there is no published way to identify them.
Sources
- official Apple Newsroom, Siri AI, a profoundly more capable and personal assistant powered by the next generation of Apple Intelligence, is here, 14 September 2026. Read 5 October 2026.
- platform documentation Apple Support, About Applebot. Search, training and AI-answer controls; Googlebot fallback; nosnippet. Read 5 October 2026.
- documentation Google Search Central, AI features and your website. Nosnippet and other preview controls limit what Search, including AI features, shows. Read 5 October 2026.
- live read robots.txt files of 30 news and business domains, listed in the methodology, requested 5 October 2026. Original measurement.