Removing Sensitive Content from the Internet
Note: The Spanish version of this resource is currently still the 2022 version. The Spanish version of the 2026 version is coming soon!
Nota: La versión en español de este recurso sigue siendo la de 2022. ¡La versión en español correspondiente a 2026 estará disponible próximamente!
So much personal information can be found online. Sometimes, it can be found on host websites (where the content is originally posted). These could be a personal blog, a social media site, or a company’s website. But even after updating privacy settings on your social media accounts, removing your information from data brokers, or taking other steps to remove the content from host websites, there can be a disconnect. The information may be removed from the sites themselves, but still appear in the search engine (e.g. Google or Bing) preview. It can include text or images. Sometimes the text no longer appears in search engine previews but images do. Or, the text may appear in popular artificial intelligence (AI) tool results. This guide will walk you through the process of removing your personal information from websites, search engines, and AI tools.
This guide is primarily intended for adults (people at least 18 years old) in the United States. If you are a minor (under 18), you can get help from the National Center for Missing and Exploited Children (NCMEC) in removing intimate images of yourself from the Internet. Google will also remove most images of you, whether they are intimate or not, from search results, if you request. You can learn how to make a request here.
Note: If you are using a VPN, and this table of contents or any other “accordion” on this page does not work properly, try switching your VPN server or temporarily turning off the VPN.
Table of Contents
How to Use This Guide
This guide begins with some necessary information about search engines for readers who may be less familiar with how they work. If you understand search engine basics, you can skip to the sections on content removal.
The sections on content removal include:
Removing content from the website where it is hosted.
Removing content from search engine results.
Removing content from the answers given by AI tools in response to questions.
If you aren’t sure about the difference between the first two, read the technology basics section.
Technology Basics
Search Engine Basics
As you read, remember the difference between a host website and a search engine like Google or Bing. Examples of a host website could be someone’s personal website, a new site, or a social media site. A search engine helps people find relevant websites based on their searches. To show the difference, here is an image of search results for “weather report new york city”:
And here’s what the host site – the Weather Channel’s website – looked like when the link was clicked.
Notice the differences between the site itself, and the search result preview shown by Google. The site claimed a 1% chance of rain, winds in a southeastward direction, and a temperature high of 79 degrees. The preview, however, claimed a 2% chance of rain, winds in an eastward direction, and a temperature of 81 degrees for the same day. When these differences exist, search engine companies call this “outdated” content, in that it no longer reflects what is on the host site. The reason why this happens is explained in the next section, but the key points are that search engines and host sites are different – the search engine is like a librarian who helps you find the book you’re looking for but is not the author of the book, while the host site is like the book itself – and that the site previews provided by search engines may show content that is no longer on the host sites.
You can enter text into a search engine to search for related websites, as in the previous example. You can also go to the image search part of a search engine (such as Google Images) and enter text or upload an image to search for related images, which is especially important if you think someone may have put sexual, intimate, or other harmful images of you online. This is the beginning of the search results for “weather report new york city” on Google Images (notice that the top results are different than for the earlier search):
Notice the differences between the site itself, and the search result preview shown by Google.
The site claimed a 1% chance of rain, winds in a southeastward direction, and a temperature high of 79 degrees.
The preview, however, claimed a 2% chance of rain, winds in an eastward direction, and a temperature of 81 degrees for the same day.
When these differences exist, search engine companies call this “outdated” content. It is outdated because the preview no longer matches what is on the host site. You can read about why this happens in a later section of this document. But the key points are that search engines and host sites are different.
The search engine is like a librarian who helps you find the book you’re looking for in the library. It is not the author of the book, so it can’t change what’s in the book. There may be a process for asking the library to remove a book or buy a new version of a book, but the librarian can’t choose to do that on their own.
The host site is like the book itself. It updates any time its author changes it. For a real book, this might mean the author published a new version. It is easier to change a website than publish a new version of a book, so authors can update websites more often. Search engines may not have the latest version of a site in their “library” yet. So they might show previews based on the old content that is no longer on the host sites.
You can search with text, as in the previous example. You can also go to the image search part of a search engine (such as Google Images) and enter text or upload an image to search for related images. This is especially important if you think someone may have put sexual, intimate, or other harmful images of you online. This is the beginning of the search results for “weather report new york city” on Google Images. Notice that the top results are different than for the earlier search. That’s because the earlier search looked for all types of content, while this one only looked for images.
Artificial Intelligence (AI) and Search Engines
Popular generative AI tools (tools that output text or images in response to a prompt, such as ChatGPT) are usually trained using online data. This means that if someone asks an AI tool about you, it may show them information about you that is online. In many ways, this is similar to a search engine showing results about you. However, the AI tool may make the information easier to find and read.
Every time a generative AI tool’s model is trained, it “learns” the data used to train it. The data becomes part of what it “knows” about the world. It responds to you by using this “knowledge” to guess what should come next as it creates the response. If it is fully retrained, that knowledge will be replaced. But the model is only occasionally fully retrained. The rest of the time, its knowledge only changes bit by bit in response to new data.
Why Would Search Engines Show Outdated Site Previews?
It can be stressful and confusing to do all the work to remove your information or images from websites, only to have them still show up in search results. Why is this happening?
Search engines have caches, which store copies of web pages. These copies help the search engine provide results more quickly. The preview text you first see in the list of search results are sometimes pulled from these caches.
Even if a website immediately removes content, it can take days or weeks for the search engine to store a newer copy of the site. The previews aren’t updated until then. During that time, the information is no longer on the actual page of the host site. But a preview of the old version (possibly containing sensitive information that you were trying to remove like an address or employer) will still show up in search results.
How Do I Remove Content from the Website That It Is On?
The first thing to consider is the type of content. Is the content intimate images of you posted without your permission? This is called nonconsensual intimate imagery (NCII) or nonconsensual distribution of intimate imagery (NDII), but you may hear it called “revenge porn”. If it is, then there are different options available to you.
It can be upsetting to come across personally identifying or intimate content online, and to figure out what to do about it. This section provides a quick step-by-step guide to removing information from the Web.
Figure out where the information/images are being stored online. This means specific pages of specific websites. In the previous example, the website is the Weather Channel’s site. The page is the New York City 10-day forecast page.
Request that the website remove the information/images. You can do this yourself or have someone like an attorney or advocate do it for you. If someone has shared sexual or intimate images of you online without your consent, see the section below on intimate content.
If the website is a data broker, see our resource on removing information from data brokers.
If the website is an app or social media site, use its reporting features. StopNCII.org has a list of common apps and how to report to them. For others, try searching for the platform name and “how to report a post.”
If the website is some other type of site, check whether it lists contact information for its webmaster (person who maintains the site) or has a contact form. Make your request directly. Google provides a helpful explanation of how to do this.
Intimate Content Posted Without Your Permission
If someone has shared sexual or intimate images of you online without your permission, the Cyber Civil Rights Initiative may be able to help. If someone is threatening to share images online and you are hoping to prevent that, visit StopNCII.org.
The Take It Down Act (TIDA) set new standards for the US around tech platform removal of NDII. It doesn’t matter whether the images are real, edited, or fake. The Federal Trade Commission has a portal with a lot more information. This includes answers to common questions.
You can start by reporting the NDII to the platform (for example, an app or website), using the platform’s reporting process. If you are not sure how to use the platform’s reporting process, searching their site or using a search engine may help you find it. Once you report to the platform, they have 48 hours to remove it and any duplicates of it from their platform. For example, TikTok would need to remove the NDII from TikTok, but not from Instagram or from Google search results.
If you are in the US, you can report a company to the FTC (Federal Trade Commission) if you think they are violating TIDA. This may be the case if:
You cannot find a reporting process after a sincere effort, or the reporting process is broken.
The host platform’s owner does not remove the information/images in 48 hours after you request them to.
Remember that the FTC cannot remove images. The purpose of reporting a company to the FTC is to report violations of TIDA. For getting the images removed, it is important to try to find and use the platform’s reporting process, not the FTC’s. If you cannot find it, see if a friend can help you.
Other Personally Identifying Content
If the host website’s owner does not remove the information/images, check if the information/images fall into categories that Google or Bing will remove from their search results. If so, make a report (Google; Bing). Many people rely on search engines to help them find content. If Google or Bing remove it from search results, this will make it more difficult for people to find even though it is still on the host site.
Both Types of Content
If the website owner does remove the information/images, search for yourself again on different search engines. Check whether that content is still visible in search result previews (like the outdated information about the weather was in the earlier example). If this is the case, go to the outdated results section of this guide.
3. You might not have found any sensitive content. Or you may be repeating this process to find and remove more results. Even in those cases, remember to search yourself regularly or use Results About You in case you need to repeat this process. If this is too stressful, think about whether there is someone in your life who could help you with it.
Image Removal Services
There are free and paid services that may help you prevent or remove either intimate images (NDII), or both intimate and other types of images.
Some free tools:
StopNCII.org is a valuable and free option for preventing NDII of you from being distributed.
The Cyber Civil Rights Initiative may be able to help you get NDII removed.
Paid services are more likely to cover a wider variety of images. Consider these questions if you are considering a paid service:
What personal information does the service company store or keep? For example, this could be names, email addresses, or phone numbers.
How does the removal process for NDII work?
What are their privacy and data security practices?
If you have done the earlier steps and content is still appearing in search history or previews, read on to learn how to address this.
How Do I Remove Content from Search Engine Results?
The first part of this section is about removing outdated results. As an example, if you had your home address removed from a website, but the website’s preview in Google Search still showed it, this would be an outdated result.
The second part of this section is about removing results that are not outdated. An example of this would be if you could not get a website owner to remove your home address. Then the search result preview is not outdated. But you might still not want it to show up in searches of your name.
How Do I Remove Outdated Search Engine Results?
Many major search engines provide tools to request removal of outdated content. This handout goes over your options for the most common search engines: Google, Microsoft Bing, Yandex, DuckDuckGo, and Yahoo. Read more about how to remove the information from data brokers.
- Create a new account.
- Use a device that the abusive person doesn't have access to.
-
You need to be logged in to a Google account to request outdated content removals. Google saves records of your requests. Anyone logged into your account will be able to see the requests. If you are worried that someone may have access to your Google account, you can create a new Google account. Use a username that is not similar to the ones you usually use, and provide as little information in the profile as possible. Or you can follow the Clinic to End Tech Abuse’s step-by-step guide to addressing Google account privacy and security issues. It is generally safer to:
Use a device that the abusive person has never had access to.
Log in through a web browser in private mode.
Once you are logged into a Google account, you can request removal of a page or an image, or make separate requests for each. Follow Google’s Remove Outdated Content instructions.
-
Bing is Microsoft’s default search engine for the Microsoft Edge browser. It has its own removal tool for outdated cache content. You must log into Bing Webmaster Tools to use it. You can log in with a Microsoft account, a Google account, or a Facebook account. See the Google section above for a discussion of Google safety concerns and how to create a new Google account if you prefer. If you are concerned about someone having access to your Microsoft account, Microsoft has instructions on how to create a new Microsoft account. Again, it is safer to use a browser in private mode and a device that the abuser/stalker has never had access to.
This tool only works for web pages, not images. There have been reports of images appearing in Bing’s image search cache long after the images were removed from the relevant page and the page itself was removed from the cache.
For certain kinds of images, you may be able to use the Report a Concern to Microsoft tool. This tool is to report digital safety issues in Microsoft products, including Bing. You can use it to report:
· Real or AI-generated intimate or sexual images of you posted without your permission.
· Images with a visible Social Security Number, bank account number, or password.
When you report a concern, it may be helpful to state in the report that you have an urgent safety concern. This may speed up the process. If this doesn’t work, another option may be to contact the Microsoft privacy team in your country. You can do so at Microsoft's privacy concerns page. Selecting “I want to make a request about my personal data” followed by “I want to contact the Microsoft privacy team or the Microsoft Data Protection Officer.” Then fill out the form.
-
The Yandex search engine is very popular among Russian speakers and in communities with large Russian-speaking populations. Yandex has a form to request removal of outdated web page content. You must be logged in to Yandex to use this form. You can log in with a Yandex account or use a Google, Facebook, or X account. If you are concerned about someone accessing your account, consider creating a new account with minimal profile information and a username that you don’t normally use. You can use this account only for making your requests.
There is also a separate form, which does not require logging in to Yandex, for requesting removal of outdated images.
-
The DuckDuckGo search engine is popular among privacy and security enthusiasts. It does not track users, store search history, or show personalized results. It does not have a built-in way to request removal of outdated content. However, it gets many of its image and web page link results from Bing. So using Bing’s content removal tools will usually remove much of it from DuckDuckGo as well.
DuckDuckGo also gets information from specialized search sites (for instance, MusixMatch for music), as well as finding them on its own. It uses this information to provide “Instant Answers.” Removing these search results may require getting the information removed from the websites storing it.
-
Yahoo’s search results are currently powered by Bing. In order to get outdated content out of Yahoo search results, follow the process described above for Bing.
How Do I Remove Sensitive Information from Search Engine Results?
This section has guidance for removing sensitive information about you from several popular search engines. The information does not need to have been removed from the original website first. This guidance was written in July 2026. Companies’ rules and processes can change at any time.
-
Google has several tools for removing certain kinds of pages from Google search results. These will not remove the content from the host websites. However, someone will not be able to find it with a Google search.
Results About You can help you figure out what information about you already appears in Google search results. This can help you find results that you want to remove.
Google has a special process to request removal of nonconsensually-posted explicit images of you from search results. This includes both real and AI-generated images. Google has a page with detailed instructions on how to do this.
Google will remove certain other types of information about you from search results. They have a page explaining what information they will remove. The page has a button that will let you start a request.
To use any of these tools, you will have to be logged into a Google account. Google saves records of your requests. Anyone logged into your account can access them. If you are worried that someone may have access to your Google account, you can create a new Google account. You may want to pick a username that is not similar to the ones you usually use. You may also want to put as little information in the profile as possible. Consider only logging into it from a device that the abuser/stalker does not know about or cannot access.
-
Microsoft has a tool that lets users report a safety concern. Since Bing is a Microsoft product, you can use this for Bing results. You will need the links that you are reporting. According to a support page, Bing “limits removal of search results to a narrow set of circumstances and conditions to avoid restricting Bing users' access to relevant information.” If the content at the link that you are reporting is illegal, make sure to say so in your report.
-
Yandex has two separate reporting tools.
The first is to report explicit content that was published without your permission. This could be real images of you, synthetic (“deepfake”) images, or both.
The second is to request personal information removal. This tool requires more detail. You will have to answer questions about:
The country you live in. Since different countries have different privacy laws, Yandex may grant a request to a resident of one country while denying it to a resident of another country.
The specific information that you “believe violates your rights.” If you live in a state that has stronger data privacy laws than the US as a whole, be sure to mention that.
The reason it should be removed (for each link).
-
DuckDuckGo does not have a built-in way to request removal of links from results. However, it gets many of its image and web page link results from Bing. If you follow the instructions for Bing, this will also help with DuckDuckGo.
DuckDuckGo provides “Instant Answers” by using links from specialized search engines (such as MusixMatch for music). If these have your personal information, you may need to get the information removed from the website storing it.
-
Yahoo currently gets its results from Bing. To remove links from Yahoo search results, follow the instructions for Bing.
How Do I Remove Content from AI Tool Responses?
Dealing with outdated AI results can be tricky. This is because generative AI tools that learn from online information already “know” the previous information. Often, the models that they use are only retrained every few months (or even less frequently). However, there are still steps that you can take.
Select an AI tool from the menu to go straight to the section for that tool:
ChatGPT
If you find results about you using ChatGPT, you can ask for them to be removed. You do not need a ChatGPT account.
Go to the OpenAI Privacy Portal. Click on the “Make a Privacy Request” button. OpenAI will ask whether you have a ChatGPT account.
Press the button that applies to you. Then select “Remove personal data from ChatGPT responses.” The first image below shows the options screen for someone with a ChatGPT account. The second shows the options screen for someone without a ChatGPT account.
If you have a ChatGPT account, you will then be directed to a login screen. If you don’t, you will be asked for your email address. This is to verify your identity and begin the process.
If you prefer video demonstrations, see the videos below. The first is for if you do not have an account. The second is for if you do have an account.
Screen readers: This silent instructional video shows how to request that your information be removed from ChatGPT if you do not have an OpenAI Account. Go to https://privacy.openai.com/policies/en. Click the “Make a Privacy Request” button in the upper right corner. This brings up a pop-up. Click the “I don’t have a ChatGPT account” button. This displays a “Your Privacy Controls” popup. Click the “Remove personal data from ChatGPT” button.
Screen readers: This silent instructional video shows how to request that your information be removed from ChatGPT if you have an OpenAI Account. Go to https://privacy.openai.com/policies/en. Click the “Make a Privacy Request” button in the upper right corner. This brings up a pop-up. Click the “I have a a consumer ChatGPT account” button. This displays a “Your Privacy Controls” popup with five option buttons. Click the “Remove personal data from ChatGPT” button. This will take you to an OpenAI account login screen.
Meta AI
You can request that Meta stop using your personal information from non-Meta sources in its AI. Meta may or may not grant your request. They review requests “consistent with your local laws.” There is not, as of June 2026, a way to make this request about information from Meta products like Facebook and Instagram.
To make your request, you will need to know what prompt led Meta AI to give out your personal information. If your information was found by an abuser, you may not know what prompt they used. In this case, try asking Meta AI questions about yourself. Save the prompts that you used and take screenshots of the results.
You do not need an account with any Meta app to make your request. Use the form on Facebook. The form explains more about what data Meta uses with Meta AI, and what Meta can and cannot do regarding your data.
Gemini / Google Search Summaries
Gemini is Google’s major generative AI tool. It is also used to create AI overviews of Google search results. Since Gemini results appear at the top of Google searches, this makes removal of personal information from it a priority for many people.
To remove your information from Gemini results, follow the steps in previous sections to:
Google then needs to grant your requests. Depending on the type of request, this could take between a day and a few weeks. After Google grants your requests, Gemini will start to “forget” your information. How long this takes depends on the websites that your information was on. This is because Google keeps information from “high-authority” websites stored longer than it does for other websites. The “forgetting” could take days to weeks once the original requests are granted.
Claude
Anthropic, the company that makes Claude, has a form for reporting content. As of June 2026, the form is not available in Spanish. In the Spanish version of this guide, we have added translations to images that illustrate the reporting process. There is no guarantee that Anthropic will grant any given request. The content must violate a law or Anthropic policy.
The reporting form asks which type of report you wish to make. In this case, select “I would like to report content for another reason.”
Anthropic will then ask for more details. It offers a list of content types to choose from. You can only choose one. Select the one that best describes the reason for your report. If the situation involves content that is illegal, sexually explicit (such as an intimate image of you), or endangering a child, it may be more effective to select whichever of those applies rather than “Private/confidential information.” However, you should tell the truth – if none of these apply, do not select them.
Once you select a category, you will be shown a new form. This form has many questions about the content as well as detailed instructions. Once you answer the questions, pressing “OK” will submit the form.
Screen Readers: This screenshot’s general structure is described in alt-text. Menu options here due to length: A. Copyright infringement; B. Trademark infringement; C. Court Orders; D. URL Blocklist; E. Defamation; F. Child Safety/CSAM; G. Fraud; H. Sexually explicit content; I. Violence and/or hate speech; J. Illegal goods; K. Private/confidential information; L. Violation of terms or usage policy; M. Other illegal or unlawful content
Grok
Grok is developed and maintained by the social media company X (formerly Twitter). It is trained on X posts as well as the Internet more generally. If you can find posts on X with your private information or harassing material, start by reporting those posts.
If the problem is that Grok is training on information that you told it in your own questions, you can set Grok to stop training on your information. This article can help you with the settings. However, Grok may still know the information that you already told it.
To report outdated information or inappropriate answers from Grok, you will need an account. Underneath the Grok result that you want to report, there should be several icons. One of those is a set of three dots. Click or press this icon. Then select “Report Issue.”
Once you select “Report Issue,” Grok will show you a menu. This menu allows you to choose what kind of problem you are reporting. If the specific problem is not an option, choose “Response feedback.” There is a text box below the menu. You can use this box to explain the problem in more detail. Be as specific as you can. If relevant, quote from X’s policies, such as those about private information or abuse and harassment. If the answer is drawing from outdated websites, you may want to try citing Section 9 of X’s privacy policy. Section 9 says that “Output contains factually inaccurate personal information relating to you, you can submit a correction request and we will make reasonable efforts to correct this information.”
Unfortunately, it is possible that none of these approaches will work. Fortunately, Grok draws a lot of information from live web search. This means that removing information about you from the original sites is likely to help.
