\n_In this blog you will hear directly from Corporate Vice President and Deputy Chief Information Security Officer (CISO) for AI, Yonatan Zunger, about how to build a plan to deploy AI safely. This blog is part of a new ongoing series where our Deputy CISOs share their thoughts on what is most important in their respective domains. In this series you will get practical advice, forward-looking commentary on where the industry is going, things you should stop doing, and more._<\/p>\n
## How do you deploy AI safely?<\/p>\n
As Microsoft\u2019s Deputy CISO for AI, my day job is to think about all the things that can go wrong with AI. This, as you can imagine, is a pretty long list. But despite that, we\u2019ve been able to successfully develop, deploy, and use a wide range of generative AI products in the past few years, and see significant real value from them. If you\u2019re reading this, you\u2019ve likely been asked to do something similar in your own organization\u2014to develop AI systems for your own use, or deploy ones that you\u2019ve acquired from others. I\u2019d like to share some of the most important ways that we think about prospective deployments, ensure we understand the risks, and have confidence that we have the right management plan in place.<\/p>\n
This is way more than can fit into a single blog, so this post is just the introduction to a much wider set of resources. In this post, I\u2019ll articulate the basic principles we use in our thinking. These principles are meant to be applicable far beyond Microsoft, and indeed most of them scope far beyond AI\u2014they’re really methods for safely adopting _any_ new technology. Because principles on their own can be abstract, I\u2019m releasing this with a companion video in which I work through a detailed example, taking a hypothetical new AI app (a tool to help loan officers do their jobs at a bank) through this entire analysis process to see how it works.<\/p>\n
<\/p>\n
Watch the video to see a real-life example on how to deploy AI safely<\/p>\n
We have even deeper resources coming soon intended to help teams and decision makers innovate safely that will build on this content. Meanwhile, if you want to learn about how Microsoft applies these ideas to safe AI deployment in more detail, you can learn about the various policies, processes, frameworks, and toolboxes we built for our own use on our Responsible AI site.<\/p>\n
## Basic principles<\/p>\n
What does \u201cdeploying safely\u201d mean? It doesn\u2019t mean that nothing can go wrong; things can _always_ go wrong. In a safe deployment, you understand as many of the things that can go wrong as possible and have a plan for them that gives you confidence that a failure won\u2019t turn into a major incident, and you know that if a completely unexpected problem arises, you\u2019re ready to respond to that as well.<\/p>\n
It also means that you haven\u2019t limited yourself to very specific kinds of problems, like security breaches or network failures, but are just as prepared for privacy failures, or people using the system in an unexpected way, or organizational impacts. After all, there\u2019s no surer guarantee of disaster than a security team saying \u201cthat sounds like a privacy problem\u201d while the privacy team says \u201cthat sounds like a security problem\u201d and neither team dealing with it. As builders of systems, we need to think about the ways in which our systems might fail, and plan for all of those, where \u201cthe systems\u201d includes not just the individual bits of software, but the entire integrated system that they\u2019re a part of\u2014including the people who use them and how they\u2019re used.<\/p>\n
These ideas probably sound familiar, because they\u2019re the basics we learned at the start of our careers, and are the same concepts that underlie everything from the National Institute of Standards and Technology (NIST) Risk Management Framework to Site Reliability Engineering. If I had to state them as briefly as possible, the basic rules would be:<\/p>\n
1. **Understand the things that might go wrong in your system, and for each of those things, have a plan.** A \u201cplan\u201d could mean anything from changing how the system works to reduce the impact of a risk, to making the failure of some component no big deal because the larger system compensates for it, to simply knowing that you\u2019ll be able to detect it and have the flexibility and tools to respond when it does.
2. **Analyze the _entire_ system, including the humans, for _any_ type of thing that could go wrong. **Your \u201csystem\u201d means the entire business process that uses it, including the people, and \u201cthings that might go wrong\u201d includes anything that could end up with you having to respond to it, whether it\u2019s a security breach or your system ending up on the front page of the paper for all the wrong reasons.
* Tip:__ Whether you\u2019re using AI software that you bought or building your own systems, you\u2019re always the builder of your own business processes. Apply your safety thinking to the complete end-to-end process either way.
3. **You think about what could go wrong from the day you get the idea for the project and do it continuously until the day it shuts down.** Planning for failure isn\u2019t an \u201cexercise\u201d; it\u2019s the parallel partner to designing the features of your system. Just as you update your vision of how the system should work every time you find a new use case or see customer needs changing, you update your vision of how the system might fail whenever it or the situation changes.<\/p>\n
You implement these three principles through a fourth one:<\/p>\n
4. **Make a written safety plan:** a discussion of these various risks and your plan for each. Don\u2019t forget to include a brief description of what the system is and what problem it\u2019s meant to solve, or the plan will be illegible to future readers, including yourself.<\/p>\n
If your role is to review systems and make sure they\u2019re safe to deploy, that safety plan is the thing you should look at, and the question you need to ask is whether that plan covers all the things that might go wrong (including \u201chow we\u2019ll handle surprises\u201d) and if the proposed solutions make sense. If you need to review many systems, as CISOs so often do, you\u2019ll want your team to create standard forms, tools, and processes for these plans\u2014that is, a governance standard, like Microsoft does for Responsible AI.<\/p>\n
These first four rules aren\u2019t specific to AI at all; these are general principles of safety engineering, and you could apply them to anything from ordinary cloud software deployments to planning a school field trip. The hard part that we\u2019ll cover in later materials is how best to identify the way things could go wrong (including when radically new technologies are involved) and build mitigation plans for them. The second rule will repeatedly prove to be the most important, as problems in one component are very often solved by changes in another component\u2014and that includes the people.<\/p>\n
Watch the video for guidance for safe AI deployment<\/p>\n
## AI-specific principles<\/p>\n
When it comes to building AI systems, we\u2019ve uncovered a few rules that are exceptionally useful. The most important thing we\u2019ve learned is that error is an intrinsic part of how AI works; problems like hallucination or prompt injection are inherent, and if you need a system that deterministically gives the right answer all the time, AI is probably not the right tool for the job. However, you already _know_ how to build reliable systems out of components that routinely err: they\u2019re called \u201cpeople\u201d and we\u2019ve been building systems out of them for millennia.<\/p>\n
The possible errors that can happen in any analysis, recommendation, or decision-making step (human or AI) are:<\/p>\n
* **Garbage in, garbage out** , also known as GIGO\u2014if the input data is bad, the output will be, too.
* **Misinterpreted data** \u2014if the data provided doesn\u2019t have _exactly_ the same meaning as the analysis expects, situations where they differ can cause subtle but dangerous errors. For example, if analysis of a loan applicant received a number it thought was \u201cmean duration of continuous employment over the past five years,\u201d but was actually receiving \u201cmean duration of each job over the past five years,\u201d it would produce extremely wrong results for consultants and other people who stack short-term jobs.
* **Hallucination** , also known as false positives\u2014the analysis introduces information not supported by the grounding data.
* **Omission,** also known as false negatives\u2014the analysis leaves out some critical caveat or context that changes the meaning of the data.
* **Unexpected preferences** \u2014every summary or recommendation chooses some aspects of the input to emphasize and prioritize over others (that\u2019s the whole point of a summary); are the factors it prioritizes the ones you wanted it to?<\/p>\n
We can combine these to add some AI-specific rules:<\/p>\n
5. **Reason about the safety of AI components by imagining \u201cwhat would happen if I replaced the AI with a room full of well-intentioned but inexperienced new hires?\u201d** Don\u2019t think of the AI like a senior person\u2014think of it like a new hire fresh out of school, enthusiastic, intelligent, ready to help, and occasionally dumb as a _rock_. Build safety into your process by considering what you\u2019d do for humans in the same place\u2014for example, having multiple sets of (AI or human) eyes on key decisions. This doesn\u2019t mean \u201chuman in the loop\u201d at every stage; instead, find the moments where it would make sense for more experienced eyes to step in and check before proceeding.
6. **Expect testing to take much more of your time, and coding to take less of your time, than with traditional software.** It\u2019s very easy to build AI software that works right in the two cases that you thought of, but much harder to make sure it works right when real user inputs are involved. Build extensive libraries of test cases, including intended uses, things confused users might do, and things threat actors might do; the line between functionality and security testing will be fuzzy. In general, you should think of AI development as a \u201cprototype-break-fix\u201d cycle rather than a \u201cdevelop-test-ship\u201d cycle.<\/p>\n
And some more rules that apply to _any_ analysis, recommendation, or decision-making stage, whether it\u2019s human or AI. (This similarity goes to the heart of Rule 5; it shouldn\u2019t be surprising that humans and AI require similar mitigations!)<\/p>\n
7. **Accompany decision-making criteria with a suite of test cases, validated by having multiple people evaluate the test cases per the criteria and tweaking the criteria until they agree with each other and your intent.** This is a good way to make sure that your written criteria (whether they be guidelines for human raters or AI metaprompts) are understood in line with your intentions. It\u2019s a good idea for the policy writers to provide a bunch of the test cases themselves, because things get lost in translation even between them and the engineering team; you can also use AI to help extend a list of test cases, then manually decide what the expected outputs for each should be. Having multiple reviewers independently decide on expectations is a good way to detect when your intentions weren\u2019t clear even to yourself.
8. **Monitor and cross-check decision making.** Send some random subset of decisions to multiple (human or AI) reviewers in parallel and monitor inter-rater agreement as a way of measuring if the stated criteria are clear enough to produce consistent answers. Automatically escalate disagreements, as well as \u201chigh impact\u201d cases (for example, large-value bank loan decisions) to more experienced people. Simultaneously, log carefully and monitor for the \u201crevealed preferences\u201d of your decision system, to ensure that they align with your intended preferences.
9. **Present information carefully.** Whenever information transits a people boundary\u2014whether this is AI outputs being presented to a human for a decision, or data collected by one team flowing to analysis run by another team\u2014you have a high risk of misinterpretation. Invest heavily here in clarity: in very sharp and rigorous API definition if it\u2019s machine-to-machine, or in extremely clear user experience if it\u2019s machine-to-human. After all, if you\u2019re running an expensive AI decision to help people and then the information is lost in translation, you aren\u2019t getting any value out of it at all\u2014and people will blame AI for the resulting human errors.<\/p>\n
The deepest insight is: novel technologies like AI don\u2019t fundamentally change the way we design for safety; they simply call on us to go back to basic principles and execute on them well.<\/p>\n
Safe AI deployment: Watch the video<\/p>\n
## Learn more<\/p>\n
You can find a detailed worked example of how to apply these ideas in this video. You can learn more about our responsible AI practices at our Responsible AI site, and about our best practices for avoiding overreliance in our Overreliance Framework.<\/p>\n
To learn more about Microsoft Security solutions, visit our website. Bookmark the Security blog to keep up with our expert coverage on security matters. Also, follow us on LinkedIn (Microsoft Security) and X (@MSFTSecurity) for the latest news and updates on cybersecurity.<\/p>\n
The post How to deploy AI safely appeared first on Microsoft Security Blog.\n<\/p><\/div>\n
View Advisory Details<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"Security Update News Update Information Title How to deploy AI safely Update ID MSSECURE:931F81227DA96AF4FDD880EEB1F03B86 Type mssecure Published 2025-05-29T16:00:00 Last Updated 2025-05-29T16:00:00 Security Impact CVSS Score…<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[4],"tags":[6,8,34,12,110,13,33,7,11,5],"class_list":["post-6113","post","type-post","status-publish","format-standard","hentry","category-category_news","tag-cve","tag-cvss","tag-cvss-00","tag-exploit","tag-mssecure","tag-news","tag-none","tag-security","tag-tapic","tag-vulnerability"],"yoast_head":"\n
How to deploy AI safely - zero redgem<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/zero.redgem.net\/?p=6113\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"How to deploy AI safely - zero redgem\" \/>\n<meta property=\"og:description\" content=\"Security Update News Update Information Title How to deploy AI safely Update ID MSSECURE:931F81227DA96AF4FDD880EEB1F03B86 Type mssecure Published 2025-05-29T16:00:00 Last Updated 2025-05-29T16:00:00 Security Impact CVSS Score...\" \/>\n<meta property=\"og:url\" content=\"https:\/\/zero.redgem.net\/?p=6113\" \/>\n<meta property=\"og:site_name\" content=\"zero redgem\" \/>\n<meta property=\"article:published_time\" content=\"2025-05-29T13:37:29+00:00\" \/>\n<meta name=\"author\" content=\"invoker\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"Written by\" \/>\n\t<meta name=\"twitter:data1\" content=\"invoker\" \/>\n\t<meta name=\"twitter:label2\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data2\" content=\"12 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\\\/\\\/schema.org\",\"@graph\":[{\"@type\":\"Article\",\"@id\":\"https:\\\/\\\/zero.redgem.net\\\/?p=6113#article\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/zero.redgem.net\\\/?p=6113\"},\"author\":{\"name\":\"invoker\",\"@id\":\"https:\\\/\\\/zero.redgem.net\\\/#\\\/schema\\\/person\\\/fbfeae8dfad117ac08a7621bee1a1dca\"},\"headline\":\"How to deploy AI safely\",\"datePublished\":\"2025-05-29T13:37:29+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\\\/\\\/zero.redgem.net\\\/?p=6113\"},\"wordCount\":2322,\"commentCount\":0,\"publisher\":{\"@id\":\"https:\\\/\\\/zero.redgem.net\\\/#organization\"},\"keywords\":[\"CVE\",\"CVSS\",\"CVSS-0.0\",\"exploit\",\"mssecure\",\"news\",\"NONE\",\"Security\",\"tapic\",\"Vulnerability\"],\"articleSection\":[\"category_news\"],\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"CommentAction\",\"name\":\"Comment\",\"target\":[\"https:\\\/\\\/zero.redgem.net\\\/?p=6113#respond\"]}]},{\"@type\":\"WebPage\",\"@id\":\"https:\\\/\\\/zero.redgem.net\\\/?p=6113\",\"url\":\"https:\\\/\\\/zero.redgem.net\\\/?p=6113\",\"name\":\"How to deploy AI safely - zero redgem\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/zero.redgem.net\\\/#website\"},\"datePublished\":\"2025-05-29T13:37:29+00:00\",\"breadcrumb\":{\"@id\":\"https:\\\/\\\/zero.redgem.net\\\/?p=6113#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\\\/\\\/zero.redgem.net\\\/?p=6113\"]}]},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\\\/\\\/zero.redgem.net\\\/?p=6113#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\\\/\\\/zero.redgem.net\\\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"How to deploy AI safely\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\\\/\\\/zero.redgem.net\\\/#website\",\"url\":\"https:\\\/\\\/zero.redgem.net\\\/\",\"name\":\"zero redgem\",\"description\":\"\",\"publisher\":{\"@id\":\"https:\\\/\\\/zero.redgem.net\\\/#organization\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\\\/\\\/zero.redgem.net\\\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"en-US\"},{\"@type\":\"Organization\",\"@id\":\"https:\\\/\\\/zero.redgem.net\\\/#organization\",\"name\":\"zero redgem\",\"url\":\"https:\\\/\\\/zero.redgem.net\\\/\",\"logo\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/zero.redgem.net\\\/#\\\/schema\\\/logo\\\/image\\\/\",\"url\":\"\",\"contentUrl\":\"\",\"width\":191,\"height\":188,\"caption\":\"zero redgem\"},\"image\":{\"@id\":\"https:\\\/\\\/zero.redgem.net\\\/#\\\/schema\\\/logo\\\/image\\\/\"}},{\"@type\":\"Person\",\"@id\":\"https:\\\/\\\/zero.redgem.net\\\/#\\\/schema\\\/person\\\/fbfeae8dfad117ac08a7621bee1a1dca\",\"name\":\"invoker\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/f17c01d7338e6932bcde121cf83569393df3374625d25afd62677cfb528f2e3e?s=96&d=mm&r=g\",\"url\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/f17c01d7338e6932bcde121cf83569393df3374625d25afd62677cfb528f2e3e?s=96&d=mm&r=g\",\"contentUrl\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/f17c01d7338e6932bcde121cf83569393df3374625d25afd62677cfb528f2e3e?s=96&d=mm&r=g\",\"caption\":\"invoker\"},\"sameAs\":[\"https:\\\/\\\/zero.redgem.net\"],\"url\":\"https:\\\/\\\/zero.redgem.net\\\/?author=1\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"How to deploy AI safely - zero redgem","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/zero.redgem.net\/?p=6113","og_locale":"en_US","og_type":"article","og_title":"How to deploy AI safely - zero redgem","og_description":"Security Update News Update Information Title How to deploy AI safely Update ID MSSECURE:931F81227DA96AF4FDD880EEB1F03B86 Type mssecure Published 2025-05-29T16:00:00 Last Updated 2025-05-29T16:00:00 Security Impact CVSS Score...","og_url":"https:\/\/zero.redgem.net\/?p=6113","og_site_name":"zero redgem","article_published_time":"2025-05-29T13:37:29+00:00","author":"invoker","twitter_card":"summary_large_image","twitter_misc":{"Written by":"invoker","Est. reading time":"12 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/zero.redgem.net\/?p=6113#article","isPartOf":{"@id":"https:\/\/zero.redgem.net\/?p=6113"},"author":{"name":"invoker","@id":"https:\/\/zero.redgem.net\/#\/schema\/person\/fbfeae8dfad117ac08a7621bee1a1dca"},"headline":"How to deploy AI safely","datePublished":"2025-05-29T13:37:29+00:00","mainEntityOfPage":{"@id":"https:\/\/zero.redgem.net\/?p=6113"},"wordCount":2322,"commentCount":0,"publisher":{"@id":"https:\/\/zero.redgem.net\/#organization"},"keywords":["CVE","CVSS","CVSS-0.0","exploit","mssecure","news","NONE","Security","tapic","Vulnerability"],"articleSection":["category_news"],"inLanguage":"en-US","potentialAction":[{"@type":"CommentAction","name":"Comment","target":["https:\/\/zero.redgem.net\/?p=6113#respond"]}]},{"@type":"WebPage","@id":"https:\/\/zero.redgem.net\/?p=6113","url":"https:\/\/zero.redgem.net\/?p=6113","name":"How to deploy AI safely - zero redgem","isPartOf":{"@id":"https:\/\/zero.redgem.net\/#website"},"datePublished":"2025-05-29T13:37:29+00:00","breadcrumb":{"@id":"https:\/\/zero.redgem.net\/?p=6113#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/zero.redgem.net\/?p=6113"]}]},{"@type":"BreadcrumbList","@id":"https:\/\/zero.redgem.net\/?p=6113#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/zero.redgem.net\/"},{"@type":"ListItem","position":2,"name":"How to deploy AI safely"}]},{"@type":"WebSite","@id":"https:\/\/zero.redgem.net\/#website","url":"https:\/\/zero.redgem.net\/","name":"zero redgem","description":"","publisher":{"@id":"https:\/\/zero.redgem.net\/#organization"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/zero.redgem.net\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"en-US"},{"@type":"Organization","@id":"https:\/\/zero.redgem.net\/#organization","name":"zero redgem","url":"https:\/\/zero.redgem.net\/","logo":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/zero.redgem.net\/#\/schema\/logo\/image\/","url":"","contentUrl":"","width":191,"height":188,"caption":"zero redgem"},"image":{"@id":"https:\/\/zero.redgem.net\/#\/schema\/logo\/image\/"}},{"@type":"Person","@id":"https:\/\/zero.redgem.net\/#\/schema\/person\/fbfeae8dfad117ac08a7621bee1a1dca","name":"invoker","image":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/secure.gravatar.com\/avatar\/f17c01d7338e6932bcde121cf83569393df3374625d25afd62677cfb528f2e3e?s=96&d=mm&r=g","url":"https:\/\/secure.gravatar.com\/avatar\/f17c01d7338e6932bcde121cf83569393df3374625d25afd62677cfb528f2e3e?s=96&d=mm&r=g","contentUrl":"https:\/\/secure.gravatar.com\/avatar\/f17c01d7338e6932bcde121cf83569393df3374625d25afd62677cfb528f2e3e?s=96&d=mm&r=g","caption":"invoker"},"sameAs":["https:\/\/zero.redgem.net"],"url":"https:\/\/zero.redgem.net\/?author=1"}]}},"_links":{"self":[{"href":"https:\/\/zero.redgem.net\/index.php?rest_route=\/wp\/v2\/posts\/6113","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/zero.redgem.net\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/zero.redgem.net\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/zero.redgem.net\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/zero.redgem.net\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=6113"}],"version-history":[{"count":0,"href":"https:\/\/zero.redgem.net\/index.php?rest_route=\/wp\/v2\/posts\/6113\/revisions"}],"wp:attachment":[{"href":"https:\/\/zero.redgem.net\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=6113"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/zero.redgem.net\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=6113"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/zero.redgem.net\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=6113"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}