Is Microsoft Copilot safe for kids?
Under our published methodology, Microsoft Copilot rates strongly on under-18 data practices: its privacy FAQ states signed-in under-18s are excluded from AI model training and from personalised advertising (checked 2026-07-28). Its terms set a 13-plus minimum with under-18 feature limits, but age rests on self-declared details, and a 2024 image-generation incident drew an FTC complaint letter before Microsoft tightened its guardrails.
Verified 28 July 2026 · An assessment by Halo Aware under our published methodology
How it scored
Each dimension is scored 0–4 against published anchors and must cite the evidence below. The overall score is weighted arithmetic, not judgement calls.
| Dimension | Score | Evidence |
|---|---|---|
| Content safety×25 | 3/4 | e7, e8, e9, e10, e11Default moderation, prompt/response blocking and abuse monitoring are documented; but the 2024 Copilot Designer episode showed filter gaps, and we found no published Copilot policy describing self-harm or crisis handling in chat. |
| Age controls×20 | 3/4 | e1, e2, e313-plus terms, under-13 block or parental consent at Microsoft-account level, and distinct under-18 feature limiting. Age rests on self-declared birth details and the terms contemplate signed-out use, so short of meaningful age assurance. |
| Data practices×20 | 4/4 | e4, e5, e6Clear policy, training opt-out, 18-month default retention, per-conversation and full-history deletion, plus a minor-specific training exclusion. That exclusion covers signed-in under-18s only; signed-out use is not age-identified. |
| Design pressure×15 | 3/4 | e7, e12General-purpose assistant on an open web surface; the terms bar adult and sexual content generation. Personalisation and memory features exist, and product framing could not be captured first-hand (landing page blocked our fetcher), so held at 3. |
| Parental visibility×10 | 3/4 | e12, e2An open web surface a parent-side tool can reasonably supervise; Halo Aware's shipped extension includes a Copilot capture module. We found no Copilot-specific parental controls, so not a 4. |
| Track record×10 | 3/4 | e10, e11One documented 2024 incident (image generation producing sexual/violent content, escalated to the FTC by a Microsoft engineer) with dated remediation within days. We found no court or regulator action against consumer Copilot involving minors. |
What we found
Every claim is sourced and dated. If something here is wrong or out of date, tell us and we will fix it.
e1The Copilot Terms of Use state a minimum age of 'usually at least 13, but sometimes 18 or older, depending on your country's laws' (checked 2026-07-28).
"WHO CAN USE COPILOT You need to be old enough to use Copilot – usually at least 13, but sometimes 18 or older, depending on your country’s laws."
https://www.microsoft.com/microsoft-copilot/for-individuals/termsofuse · captured 28 July 2026
e2The Copilot Terms of Use state Microsoft may turn off or limit features for users under 18, or for users not logged in, for legal or safety reasons (checked 2026-07-28).
"If you’re under 18, or if you use Copilot without logging in, we might turn off or limit some features for legal or safety reasons. If we ask for your birthday and country when you sign up or log in, you must give us your real information."
https://www.microsoft.com/microsoft-copilot/for-individuals/termsofuse · captured 28 July 2026
e3The Microsoft Privacy Statement states certain products either block users under 13 or require parental or guardian consent or authorisation before use, including at account creation (checked 2026-07-28).
"For users under the age of 13 or as specified by law in their jurisdiction, certain Microsoft products and services will either block users under that age or require parental or guardian consent or authorization before use, including when creating an account."
https://www.microsoft.com/en-us/privacy/privacystatement · captured 28 July 2026
e4Copilot's Privacy FAQ states Microsoft does not train Copilot on data from signed-in users under 18, and does not serve personalised advertising to authenticated users under 18 (checked 2026-07-28).
"We do not train Copilot on data from the following types of users: Users signed into Copilot with an organizational Entra ID account. […] Users under the age of 18 who are signed in to Copilot. Users who have opted out of model training. […] Copilot does not serve personalized advertising to authenticated users under the age of 18."
https://support.microsoft.com/en-us/microsoft-copilot/privacy-faq-for-microsoft-copilot · captured 28 July 2026
e5Copilot's Privacy FAQ states conversation activity is used for AI model training by default except for excluded categories, with an opt-out that excludes future conversations (checked 2026-07-28).
"Except for certain categories of users or users who have opted out, Microsoft uses data from Bing, MSN, Copilot, and interactions with ads on Microsoft for AI training. […] Opting out will exclude your future conversations from being used for training these AI models, unless you choose to opt back in."
https://support.microsoft.com/en-us/microsoft-copilot/privacy-faq-for-microsoft-copilot · captured 28 July 2026
e6Copilot's Privacy FAQ states conversation activity is stored for 18 months by default, and users can delete individual conversations or their entire history at any time (checked 2026-07-28).
"By default, we store conversation activity for 18 months. You can delete individual conversations or your entire conversation history at any time."
https://support.microsoft.com/en-us/microsoft-copilot/privacy-faq-for-microsoft-copilot · captured 28 July 2026
e7The Microsoft Services Agreement's Code of Conduct prohibits using the Services to generate or share content involving nudity, pornography, graphic violence or self-harm (checked 2026-07-28).
"Don’t publicly display or use the Services to generate or share inappropriate content or material (involving, for example, nudity, bestiality, pornography, offensive language, graphic violence, self-harm, or criminal activity)."
https://www.microsoft.com/en-us/servicesagreement · captured 28 July 2026
e8The Copilot Terms of Use state Microsoft may block, restrict or remove prompts or content that violate the terms, or that could lead Copilot to create a violating response (checked 2026-07-28).
"We may block, restrict, or remove your Prompts or other content from you that violates these Terms, or that could lead Copilot to create a Response that violates these Terms."
https://www.microsoft.com/microsoft-copilot/for-individuals/termsofuse · captured 28 July 2026
e9The Microsoft Services Agreement states inputs to and outputs from the AI services are processed and stored to monitor for and prevent abusive or harmful uses or outputs (checked 2026-07-28).
"As part of providing the AI services, Microsoft will process and store your inputs to the service as well as output from the service, for purposes of monitoring for and preventing abusive or harmful uses or outputs of the service."
https://www.microsoft.com/en-us/servicesagreement · captured 28 July 2026
e10Documented recordCNBC reported (6 March 2024) that a Microsoft engineer wrote to the FTC chair warning that Copilot Designer was generating sexual and violent imagery; CNBC said it reproduced his findings.
"He’s warned Microsoft of the sexual and violent content that the product, Copilot Designer, is creating, but said the company isn’t taking appropriate action. On Wednesday, Jones escalated the matter, sending letters to FTC Chair Lina Khan and to Microsoft’s board, which were viewed by CNBC."
https://www.cnbc.com/2024/03/06/microsoft-ai-engineer-says-copilot-designer-creates-disturbing-images.html · captured 28 July 2026
e11Documented recordCNBC reported (8 March 2024) that Microsoft changed its AI guardrails and began blocking prompts that had caused Copilot's image tool to create violent or sexual images, following the engineer's FTC letter.
"Microsoft begins blocking some terms that caused its AI tool to create violent, sexual images […] Microsoft has made changes to its artificial intelligence guardrails after a staff AI engineer wrote to the Federal Trade Commission of his concerns regarding Copilot's image-generation AI."
https://www.cnbc.com/2024/03/08/microsoft-blocking-terms-that-cause-its-ai-to-create-violent-images.html · captured 28 July 2026
e12Observed by Halo AwareCopilot runs as an open web app at copilot.microsoft.com; Halo Aware's shipped Chrome extension (Chrome Web Store build v1.5.0 at capture) includes a Copilot capture module, so a parent-side supervision tool can monitor conversations there. Our automated sweep on 2026-07-28 could not verify live capture: Copilot's anti-bot challenge blocked the rig's automated browser before a reply rendered.
https://copilot.microsoft.com · captured 28 July 2026
Platform operator, or spotted an error? Email alerts@haloaware.com. We review disputed claims promptly, correct what is wrong, and log corrections publicly on this page.
Halo Aware monitors Microsoft Copilot
If your child uses Microsoft Copilot in their browser, the Halo Aware extension can give you usage patterns, topic summaries and risk alerts. Summaries, not surveillance.
Get started free