Frequently Asked Questions
Get the most out of Copilot
Get the most out of Copilot
Microsoft Copilot is an AI-powered digital assistant designed to engage in conversations and help people with a range of tasks and activities on their devices. Copilot can answer questions, create images or drafts of written content, suggest different ways to word things you've written, and much more.
Copilot is like having a research assistant, personal planner, and creative partner at your side. With Copilot you can:
- Ask a question, explore a curiosity, or practice a conversation. When you ask complex questions, Copilot gives you detailed replies. Feel free to ask follow-ups, such as “can you explain that more simply?” or “please give me more options.”
- Get a real answer. Copilot looks at search results across the web to offer you a summarized response and links to its sources.
- Be creative. When you need inspiration, Copilot can help you write poems, stories, or even create a brand-new image. Copilot works best when you’re more descriptive, so get creative and add details like adjectives, locations, or even artistic styles such as ‘digital art’ and ‘photorealistic.’
Each conversation with Copilot will have a limited number of interactions to keep the conversation interesting, relevant, and grounded in search.
Talking to Copilot should feel natural. Type text or speak with your voice — whatever works best for the conversation. Copilot will automatically recognize the language you are using and respond accordingly. You can ask Copilot to speak in another language as well, but it is optimized for English.
As you’re interacting, Copilot can:
- Search the web in a way that feels natural to how you talk, text, or think. Get a single, summarized answer with a consolidated set of links, or a detailed response and sources for more complex questions.
- Ask follow-up questions to your initial question or search to get more helpful replies.
- Rephrase the response and make things easier to understand. Just say something like, “Can you explain that more simply?”
- Respond with different formats like a table or numbered list when you ask for it.
- Help you write poems, stories, speeches, or share ideas for a project.
When you engage Copilot via voice, Microsoft stores voice data in accordance with our data retention policy. Microsoft uses voice conversations and other interactions with Copilot to provide and improve the service as outlined by the Microsoft Privacy Statement. As announced in August, we are also providing our customers with advance notice of training our generative AI models in Copilot on this data, and are providing a clear and easy way to opt out of such training. We are rolling this out gradually to ensure we comply with privacy laws around the world, and are not training on consumer data from the UK, the European Economic Area (EEA), and certain other countries at this time.
You can access Microsoft Copilot at https://copilot.microsoft.com/, Microsoft Copilot Android and iOS apps, Copilot in Windows app, or via Edge by clicking the Copilot icon on the top right. To learn tips for using Copilot in Edge and other AI features in the browser, visit AI tips page.
Visual Search allows you to upload images to Copilot and ask questions about them. For example, you can ask questions about images that are difficult to describe, get a recipe for a dish you don’t know the name of, identify a dog breed you don’t recognize, and much more. To use Visual Search, click on the camera icon in the input bar in Copilot, upload a picture from your device or provide a website link, and ask away.
Tell me about Responsible AI
Tell me about Responsible AI
Copilot searches for relevant content across the web and then summarizes the information it finds to generate a helpful response. It also cites its sources, so you’re able to see links to the content it references.
At Microsoft, we take our commitment to responsible AI seriously. Copilot is developed in accordance with our AI principles. We are working with our partner OpenAI to deliver an experience that encourages responsible use, and foundational model work. We've designed the Copilot user experience to keep humans at the center and developed a safety system designed to mitigate failures and avoid misuse with things like content filtering, operational monitoring, abuse detection, and other safeguards. Responsible AI is a journey, and we are continually improving our systems along the way. To learn more about how to use Copilot responsibly, please review our Terms of Use and Code of Conduct.
While Copilot works to avoid sharing unexpected offensive content in search results and has taken steps to prevent Copilot features from engaging with potentially harmful topics, you may still see unexpected results. We’re constantly working to improve our technology to prevent harmful content.
If you encounter harmful or inappropriate content, please provide feedback or report a concern to Copilot by clicking the ‘give feedback’ button in settings or use the flag icon below each response on mobile and the Copilot web page. On the Copilot app, you can long press the response and select “Report.” We will continue to work with people’s feedback to provide a safe search experience.
Copilot aims to respond with reliable sources, but AI can make mistakes, and third-party content on the internet may not always be accurate or reliable. Copilot may misrepresent the information it finds, and you may see responses that sound convincing but are incomplete, inaccurate, or inappropriate. Use your judgment and double check facts before making decisions or taking action based on Copilot’s responses. Reviewing Copilot’s citations is a good place to start checking for accuracy.
To share site feedback or report a concern, select the ‘give feedback’ button in settings or use the flag icon below each response on mobile and the Copilot web page. On the Copilot app, you can also long press the response and select “Report.” We will continue working with people’s feedback to provide a safe search experience for all.
Learn more about chat history
Learn more about chat history
Yes, if you are signed in, the conversations you have with Copilot are saved; We call this “Chat history”. You can view and access each conversation you previously had with Copilot. For example, if you ask Copilot for information about the first person to walk on the moon, the conversation you have will be saved in chat history, which you can return to at any time. By default, we store chat history data for 18 months from the last update to the conversation. You can access and delete your chat history by visiting Home and selecting the time machine icon.
There are two settings available in Copilot Settings to control Copilot data:
- You can turn off AI’s ability to use your data to train models. To do this, go to Sign-in > Settings > Account > Privacy >Model training. You can also go to Privacy > Export/Delete, to delete chat history from account.microsoft.com. Please note, if you are in the European Economic Area (EEA), the UK, and certain other countries Copilot is not using your data to train models currently.
- You can turn off personalization. If you use this setting, you will still have access to chat history and all future chats are saved, but Copilot experiences are not personalized. If you choose to enable the experience, Copilot also uses Bing search history and expressed interest from MSN to create more personalized, fluid conversations. For users in the EEA, the UK, and a limited number of other countries, we are evaluating options before offering this level of personalization for those users. Where offered, Copilot will notify users and provide them with a clear and easy way to opt-out of personalization.
To delete your Copilot chat history, from Settings go to Account > Privacy > Export or delete history, to delete your history from your Microsoft Privacy Dashboard. In addition, in your Microsoft Privacy Dashboard click ‘Browsing and search’ and scroll down to the Search history section. In the search box search for and delete all instances that include “Microsoft Copilot”.
If you are wanting to delete chats from Copilot before 10/01/2004 (legacy experience), in your Microsoft Privacy Dashboard click ‘Browsing and search’, scroll down to the Copilot activity history section and click ‘Clear all Copilot activity history and search history’.
Yes, we keep some of your chat history including prompts, to provide and improve the service. Your prompts are maintained in accordance with our Privacy Statement and used to monitor performance, troubleshoot problems, diagnose bugs, prevent abuse, and other product performance analytics necessary to provide and improve Copilot. Only the data visible in chat history in product is used for personalized answers.
To understand more about Microsoft Privacy Practices, review the Microsoft Privacy Statement and our privacy center.
Personalize your Copilot experience
Personalize your Copilot experience
If you turn on personalization, we will use your recent conversations and Microsoft activities to personalize your experience. Your Microsoft activities may, depending on your privacy settings, include your Bing search activities, MSN activities, and inferred interests. If you do not want us to personalize using a specific conversation from your chat history, search history or inferred interests, you can delete them on the Microsoft Privacy Dashboard. You can turn off personalization at any time by clicking your account icon at top right corner of the page > account > privacy > personalization. On Edge, you can turn off by clicking on “...”: from Copilot setting menu > permissions and privacy > personalization.
Allowing Copilot personalization means you’re letting Microsoft use your recent conversations and Microsoft activities including your Bing search and MSN activities to make Copilot more personalized for you. Copilot remembers your history and may make inferences from that history about your interests, hobbies, etc., to provide personalized experiences. However, as a Copilot user, you are able to delete all chat messages.
No. The Copilot personalization toggle (available on all Copilot surfaces), controls whether recent conversations and your Microsoft activities are used for personalization. The Edge personalization and advertising toggle controls whether Edge browsing activities can be used across Microsoft surfaces for personalization including Copilot.
You control which chats are used to personalize your experience. If you don’t want Copilot to personalize using your chat history, you can clear Copilot activity history on Microsoft Privacy Dashboard. Once deleted, the chats will not be used to personalize experiences.
Personalization is on by default except for those in the European Economic Area (EEA), the UK, and certain other countries. Users have the option to opt out of the experience. In your conversations, you will be notified whether your experiences are personalized or not. We will display a notice like " Copilot can learn about your interests and personalize your experience.” if personalized experience is allowed. A notice such as “Conversations are used to train AI and Copilot can learn about your interests.” will be displayed when both training and personalization is allowed. Depending on your region, personalization may be available by default. Personalization is not available for unauthenticated users.
We’re following our Responsible AI practices and rolling out this feature slowly; as a result, the feature may not yet be available in the European Economic Area (EEA), the UK, and certain other countries.
You can turn off personalization through the personalization setting in Copilot’s settings menu.
If you change your personalization consent from “yes” to “no,” we will only stop providing a personalized Copilot experience. We will keep your chat history so you can still view your past conversations. However, your future Copilot experience will no longer be personalized.
Yes, you can turn off personalization within Copilot settings at any time. Turning off personalization won’t delete your chat history.
Regardless of your user settings, Copilot is designed not to personalize interactions based on some topics to protect your privacy and to prevent the use of potentially sensitive information. For example, we take steps to exclude context from your past conversations that may be considered sensitive, even if chat history and personalization are turned on in your settings. Users should use caution when sharing information they consider to be sensitive with Copilot.
Review data use and control options
Review data use and control options
Generative AI refers to a category of AI models that analyze data, find patterns and use these patterns to generate or create a new output, such as text, photo, video, code, data, and more. “Training” an AI model means providing it with data to help it learn to make predictions or decisions. Training is a broad concept that includes many different activities to help models learn pre-training a generative AI model, fine tuning a model, or training a classifier or filter to help models provide more appropriate results.
These models use patterns and correlations learned from their training data to predict what comes next in a sequence. They are tuned to learn general relationships in language, not to memorize specific segments of training data. They do not store or have access to the original training data. Instead, generative AI models are designed to generate new expressive works and content. We also take additional steps to prevent these models from inadvertently reproducing their training data, such as conducting testing and building filters that screen out previously published or used material.
Microsoft uses publicly available data, mostly collected from industry-standard machine learning datasets and web crawls, like search engines. We exclude sources with paywalls, content that violates our policies, or sites that have opted-out of training using web controls that we published. On top of this, we do not train on data from domains listed in the Office of the United States Trade Representative (USTR) Notorious Markets for Counterfeiting and Piracy list.
We do not train on data from our commercial customers, or any data from users logged into an organizational M365/EntraID account. We also do not train on data from M365 personal or family subscriptions or on data from users who are not logged into their Microsoft Account.
Microsoft will begin to use consumer data for AI training in certain countries later this year. For users logged into a Microsoft account, you will be able to control whether your consumer data is used for training the generative AI models used in Copilot and other AI offerings from Microsoft through an opt-out control. Opting out will exclude your past, present, and future consumer data from being used for training these AI models, unless you choose to opt back in. This setting will be made available across web, desktop, and mobile versions of all in-scope consumer products over the coming months. You will also see in-product messages notifying you about a new user control, including on Bing, MSN, and Copilot.
We will limit the data we use for training. We will not train on data from our public sector or enterprise customers, unless they expressly opt-in, nor will we train on any data from M365 consumer offerings. Learn more here. We also will not train AI models on personal account data like your Microsoft account profile data or email contents. If any images are included in your AI conversations, we take steps to de-identify them such as removing metadata or other personal data and blurring images of faces.
Your personal interactions with our services like queries and responses are kept private and are not disclosed without your permission. You can adjust your personalization in settings.
We remove information that may identify you, like names, phone numbers, device or account identifiers, sensitive personal data, physical addresses, and email addresses, before training AI models.
Your data remains private when using our services. We will continue to evaluate our models and take steps to prevent models from reproducing user information or past conversations. We will protect your personal data as explained in the Microsoft Privacy Statement and in compliance with privacy laws around the world.
We do not use data from users under 18. We only train our generative AI models on data from MSA authenticated users 18 years or older. In addition, if you are in the European Economic Area (EEA), United Kingdom, and certain other countries listed below, your data is not being used to train AI models at this time.
Microsoft will begin to use data from Bing, MSN, Copilot, and interactions with ads on Microsoft for AI training later this year. This includes anonymous search and news data, interactions with ads, and voice and text conversation data with Copilot like prompts, queries, chats, responses, and other. This data will be used to improve Copilot and our other generative AI models to create a better user experience for you and others. These generative AI models may also be used to improve other Microsoft products and services. For users logged into their Microsoft account, you are able to control whether your consumer data is used for training generative AI models used in Copilot and other AI offerings from Microsoft through opt-out control.
By using real-world consumer data to help train our underlying generative AI models, we can improve Copilot in the future and offer a more personalized and relevant natural language experience. For example, our AI models can use these conversations to learn from human communication patterns, like understanding colloquial phrases or local references. The more diversity in conversations our AI models are exposed to, the better they will understand and serve important regional languages, geographies, cultural references, and trending topics of interest to users.
We also use Copilot conversations and other consumer data for digital safety, security, and compliance purposes as outlined in the Microsoft Privacy Statement.
Before we start training, we’ll also make it easy for people to opt out, if they wish, of having the data they share in these services used to train our generative AI models. We will provide opt-out controls starting in October. To give users ample time to consider their choice, we also won’t begin training AI models on this data until at least 15 days after providing these opt-out controls. At this time, we will not use consumer data to train models in the European Economic Area (EEA), the United Kingdom, and certain other countries listed below. We will roll this out to those regions gradually, in order to ensure we get this right for consumers and comply with local privacy laws around the world.
This opt-out setting will not exclude your data from being used for other general product or system improvements nor from use for advertising, digital safety, security, and compliance purposes as outlined in the Microsoft Privacy Statement.
Consumers who are 18 years or older and logged into their Microsoft (consumer) accounts will be offered the option to opt-out of AI training starting in October. This setting will not be available in the following markets in September: Austria, Belgium, Brazil, Bulgaria, Canada, China (including Hong Kong), Croatia, Cyprus, the Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Iceland, Ireland, Israel, Italy, Latvia, Liechtenstein, Lithuania, Luxembourg, Malta, the Netherlands, Norway, Nigeria, Poland, Portugal, Romania, Slovakia, Slovenia, South Korea, Spain, Sweden, Switzerland, the United Kingdom, and Vietnam. This includes the regions of Guadeloupe, French Guiana, Martinique, Mayotte, Reunion Island, Saint-Martin, Azores, Madeira, and the Canary Islands. This means that AI offerings will be available in those markets, but no user data will be used for generative AI model training in those locations until further notice.
No, nothing you say to Copilot will ever be made public. Copilot will not have direct access to your chat history or personal data during conversations with other users.
Yes, some Copilot conversations are subject to both automated and human review for both product improvement and digital safety purposes. For example, to build, train, and improve the accuracy of our automated methods of processing (including artificial intelligence or AI), we manually review some of the output produced by the automated methods against the underlying data. We include human feedback from AI trainers and employees in our training process. For example, human feedback that reinforces a quality output to a user’s prompt, improving the end user experience.
We may also review conversations flagged as a violation of the Code of Conduct in the Terms. Our Code of Conduct prohibits use of the Copilot service to create or share inappropriate content or material. Some conversations are reviewed when a violation of the Code of Conduct is suspected.
Limited human review is required as part of the investigation process when a violation of the Code of Conduct is suspected. To ensure that our services are safe and secure for everyone, a complete opt-out of human review is not available.
Yes, you can opt-out of AI training and still have personalization turned on. In this case, Copilot will remember recent conversations to give you a more personalized response, but Microsoft will not use your conversations and other Microsoft activities for generative AI model training. For example, Copilot will remember that you are a vegetarian when suggesting dinner recipes.
We will not share your data with third parties or partners for AI training purposes without your permission.
Your conversations with Copilot, like queries and responses, are not disclosed or sold to third parties, except for the following limited instances:
Microsoft shares some personal data with third parties in accordance with our Microsoft Privacy Statement. For example, we may share your personal data with Microsoft-controlled affiliates and subsidiaries, with vendors working on our behalf, and when required by law or to respond to legal process.
To improve Copilot’s safety and ability to detect risks, the Copilot team partners with external research organizations to review and evaluate Copilot logs. For example, external research organizations can help review Copilot conversation logs to understand the variety of queries used to seek extremist content, compare trends across the industry, and advise techniques for better finding and mitigating harm.
Frequently Asked Questions
We’re transforming Edge for Business into a new category of secure enterprise browser…the secure enterprise AI browser.
It has Microsoft 365 Copilot woven into everyday workflows, with enterprise-ready compliance and controls, so your workforce can take advantage of new capabilities that put AI right in their flow of work.
It’s integrated deeply with your Microsoft 365 Graph, surfacing what matters, when it matters. Combined with an understanding of the web, it can streamline workflows and connects the dots across apps, tabs, and tasks. In fact, Edge for Business is the first AI browser experience that brings together public web information and your organizational data securely in one place.
And you can count on it to have the secure enterprise browser foundations for AI—and the many scenarios you already need to address today.
No, this is the same Edge for Business you already know. By enabling Copilot Mode, users can enjoy the advanced capabilities of AI browsing.
Edge for Business is built on a secure foundation. The features of Copilot Mode follow data loss protection (DLP). Agent Mode uses modern CUA (Computer-Using Agent) models to complete tasks for users in their browsers. We are excited about the many exciting and emergent possibilities this feature brings, but as a new technology, it introduces new potential attack vectors that we and the rest of the industry are taking on.
We take very seriously our responsibility to keep our users safe on the web.
The key to any protection strategy is defense-in-depth:
- Untrusted input: We start by assuming that any input from an untrusted source may contain unsafe instructions.
- Agent Mode access scoped to known sites by default – by default Edge only allow access to a curated list of sites. IT Admis can curate an allow list of sites that Copilot can interact with.
- SmartScreen protection – while Copilot is controlling the browser, suspicious or bad sites are blocked automatically by Microsoft Defender SmartScreen, and the agent is prevented from bypassing the block page.
- Additional mitigations - Edge Agent Mode uses Azure Prompt Shields to analyze whether data is malicious. Our built-in safety stack ensures that Copilot is trained specifically to detect and report safety violations if malicious content tries to encourage violence or harmful behavior. Lastly, Spotlighting (in testing) – Kiciman, et al, of Microsoft Research, described a technique called Spotlighting to better separate user instructions from grounding content (documents and web pages) so the model can better ignore injected commands without impacting efficacy. We will be testing Spotlighting with Actions in Edge and will report on its effectiveness.
- Detect deviations: Prompt injections generally cause the model to do something different from what the user asked of it. Mitigations can be created to detect and block those deviations.
- Relevance checks – give the user a chance to stop an action if a secondary model detects possible task drift.
- High risk site prompts – when a context is detected to be sensitive (e.g. email, banking, health, sensitive topics) the model will stop and request permission to continue.
- Task Tracker (in testing) – Paverd et al described a novel technique known as Task Tracker, which monitors activation deltas to detect when the model drifts from the user’s original intent after processing external data. We are integrating these techniques into our orchestration layer, validating their precision, and reducing false positives with the MAI Security team. We will report on progress here as well.
- Limit access to sensitive data or dangerous actions: Simply put, if the model can’t get to something or do something bad, then the risks are lower.
Other restrictions include (but are not limited to):
- No interaction with edge:// pages (e.g., Settings) or UI outside a tab’s web content.
- External app launches and extensions are blocked (protocol handlers).
- Downloads and uploads are disabled.
- No ability to open the file or directory selection dialog.
- No access to data or apps outside of Edge
- Context menus are disabled.
- Tab audio is muted by default.
- Site Permission changes are blocked (For example, if a site requests camera access permissions, the agent cannot grant that permission).
All new features will be available in limited preview for desktop users in coming months, with mobile availability coming soon after.
We will be announcing and launching new features to Copilot Mode overtime. It will provide an opportunity to experience and test new AI browser features in Edge. IT Admins can continue to set individual policies for their tenants.
At Microsoft Ignite features will include:
- The Copilot-inspired new tab page, work cards, and widgets which will be available to all users
- Agent Mode which will be available to Copilot Mode enrolled users with the Microsoft 365 Copilot license.
- Enhanced contextual grounding which will be available to users who enroll in Copilot Mode, regardless of their Microsoft 365 license.
- Other features like Daily Briefing are coming soon!
Copilot Mode will be available in limited preview for desktop users next month, with mobile availability coming soon after. Certain features like Agent Mode will start in the Frontier program and we hope to expand access over time. Individual features like Agent Mode and Daily briefing will require a Microsoft 365 Copilot license.
AI browsing is set to boost productivity, but until now, organizations have been asked to drop their bar on enterprise-grade security to benefit from it. Microsoft aims to change this by transforming the secure enterprise browser into a secure enterprise AI browser, making AI-assisted browsing safe for workplaces.
Our approach to AI browsing delivers agentic, proactive, and contextual intelligence, helping users streamline workflows, stay on task, and connect the dots across data silos. Additionally, AI browsing for work ensures that your unique work context is integrated. Edge for Business is the first AI browser to securely combine public web data with organizational data, applying AI power to both web content and internal projects. This innovation is backed by Microsoft's commitment to security, controls, and enterprise data protection.
Copilot Mode is off by default, and in your control. IT Admins can enable and customize it in the Edge management service in the Microsoft 365 admin center. Other AI features like Copilot in Edge sidepane and contextual grounding are available as their own policies outside of Copilot Mode.
Yes, IT Admins can enable and customize Copilot Mode in the Edge management service in the Microsoft 365 admin center. IT Admins can enable and disable individual AI features like Copilot Mode, Copilot in Edge, and contextual awareness. Additionally, they can turn on/off the Agent Mode feature and customize the list of allowed sites it can access.
Users will have several visual indicators. If they have chosen to “watch progress” the tab Copilot is actively working in will have a cursor in the tab heading and a colored hue around the webpage. The user won’t be able to interact with the webpage while Copilot is active. Users will have the option to stop or pause Copilot if they would like to intervene. Additionally, the chain of thought showing the steps being taken by Copilot will be visible in the chat window., including progress in the Copilot sidepane and the browser window will have a box around it (grab consumer language for this)
You can learn more by reading this blog.
- * Feature availability and functionality may vary by device type, market, and browser version.

Try Copilot today!
Try Copilot today!