---
title: "Microsoft Copilot | Ethical Stack"
url: https://ethicalstack.ai/products/copilot/
catalog_version: f6baff2119d0e9158520da7fbe4232011dac79f25bd08757436d7a1dbd593db2
---

# Microsoft Copilot

Consumer assistant combining conversation and web-grounded answers with links to sources.

Canonical profile: https://ethicalstack.ai/products/copilot/

Scope: Microsoft Copilot: cloud access. Free consumer access; paid Microsoft 365/Copilot options vary. Ratings describe the stated product, company or configuration in each finding.

Published overall: 2; 5/6 dimensions rated.

Round the equal-weight mean of available numerical ratings to one decimal; require at least four of six for an overall score. BDS status stays outside the mean; for eligible profiles, a boycott target caps the final score at 2/10 and a pressure target subtracts one point (minimum 1). Not listed and not checked do not adjust the score.

Boycott target: overall capped at 2/10.

## Energy practices

Finding: 7/10

Assessment: https://ethicalstack.ai/products/copilot/#energy

Microsoft reports annual company-wide renewable matching, regional data-center PUE/WUE, and FY25 improvement in global WUE. Copilot consumer modes offer quick and deeper response choices, but no Copilot-request environmental footprint or serving region is published.

Reason: 1 + 2 reporting + 1 electricity + 3 controls = 7/10. This grade covers documented practices and transparency; it does not measure the energy or emissions of a request.

Argument: The company’s dated regional PUE/WUE and reported decrease in global WUE support D2 and C3 at Microsoft data-center scope. The FY25 annual renewable match supports E1 only. Copilot consumer Quick response, Think Deeper, Smart, and Study modes provide user-facing workload choices, but no product-level environmental results tie them to resource savings. Company evidence must not be mistaken for the route of a specific Copilot response.

Type: conditional; confidence: medium

Scope (dimension): Consumer Copilot app; Microsoft data-center metrics and annual renewable matching describe Microsoft infrastructure company-wide, not the Copilot request route.

Research reviewed: 2026-10-24; evidence date: 2026-10-24

Unknowns: Copilot request/model/region/facility allocation; actual energy, emissions, water and response-mode use; hourly electricity source at serving locations; hardware and construction footprint.

```json
{
  "method": "practices-transparency-v1",
  "scope": "Consumer Copilot app; Microsoft data-center metrics and annual renewable matching describe Microsoft infrastructure company-wide, not the Copilot request route.",
  "components": {
    "disclosure": {
      "score": 2,
      "reason": "Microsoft publishes dated global/regional PUE and WUE plus company emissions; the Copilot consumer app has no request-level energy/carbon/water allocation.",
      "sourceUrls": [
        "https://support.microsoft.com/en-us/microsoft-365-copilot/conversation-modes-in-microsoft-copilot",
        "https://datacenters.microsoft.com/sustainability/efficiency/",
        "https://www.microsoft.com/en-us/corporate-responsibility/topics/sustainability/report/"
      ]
    },
    "electricity": {
      "score": 1,
      "reason": "Microsoft reports annual global electricity matching with renewable energy for FY25. This is an annual company-wide match, not hourly matching or Copilot-route sourcing.",
      "sourceUrls": [
        "https://support.microsoft.com/en-us/microsoft-365-copilot/conversation-modes-in-microsoft-copilot",
        "https://datacenters.microsoft.com/sustainability/efficiency/",
        "https://www.microsoft.com/en-us/corporate-responsibility/topics/sustainability/report/"
      ]
    },
    "demand": {
      "score": 3,
      "reason": "Microsoft reports FY25 global data-center WUE improved from 0.30 to 0.27 L/kWh and documents cooling-efficiency controls. This is a measured physical resource-intensity reduction plus controls at company facilities, not a Copilot-specific saving.",
      "sourceUrls": [
        "https://support.microsoft.com/en-us/microsoft-365-copilot/conversation-modes-in-microsoft-copilot",
        "https://datacenters.microsoft.com/sustainability/efficiency/",
        "https://www.microsoft.com/en-us/corporate-responsibility/topics/sustainability/report/"
      ]
    }
  },
  "actualImpactMeasured": false,
  "meaning": "Documented environmental practices and transparency. Zero component points mean a reviewed evidence gap."
}
```

- [Conversation modes in Microsoft Copilot](https://support.microsoft.com/en-us/microsoft-365-copilot/conversation-modes-in-microsoft-copilot) (primary; checked 2026-10-24)

- [Measuring energy and water efficiency for Microsoft datacenters](https://datacenters.microsoft.com/sustainability/efficiency/) (primary; checked 2026-10-24)

- [2026 Environmental Sustainability Report](https://www.microsoft.com/en-us/corporate-responsibility/topics/sustainability/report/) (primary; checked 2026-10-24)

## Training data

Finding: Depends on the model

Assessment: https://ethicalstack.ai/products/copilot/#training-rights

Microsoft’s own documentation describes Copilot as drawing on web and personal/work data, and notes model-training information may differ between older and updated app versions. The reviewed product evidence does not fix one underlying model corpus.

Reason: Identify the current Copilot app/version and selected model route; then assess that provider model’s training corpus separately. This grades documented training-data safeguards, not a legal ruling.

Argument: This profile mixes model routes and grounded sources, and the cited training FAQ itself distinguishes old and new Copilot versions. A single score would imply a stable training corpus that the profile does not establish.

Type: conditional; confidence: low

Scope (dimension): Microsoft Copilot consumer service; model and grounded-content routes vary by app and configuration.

Research reviewed: 2026-10-24; evidence date: not recorded

Unknowns: Current default model/version for this user-facing Copilot app and its model-specific source/rights disclosures need to be fixed.

```json
{
  "method": "training-rights-evidence-v1",
  "scope": "Microsoft Copilot consumer service; model and grounded-content routes vary by app and configuration.",
  "components": null,
  "routeDependency": "Identify the current Copilot app/version and selected model route; then assess that provider model’s training corpus separately.",
  "modelReference": null,
  "meaning": "Documented training-data rights safeguards. Low points may reflect missing evidence; they do not establish infringement."
}
```

- [Privacy FAQ for Microsoft Copilot](https://support.microsoft.com/en-us/microsoft-copilot/privacy-faq-for-microsoft-copilot) (primary; checked 2026-10-24)

- [What information does Copilot use to answer my prompt?](https://support.microsoft.com/en-us/microsoft-365-copilot/what-information-does-copilot-use-to-answer-my-prompt) (primary; checked 2026-10-24)

## Your data & privacy

Finding: 5/10

Assessment: https://ethicalstack.ai/products/copilot/#privacy

Consumer conversation activity can be used for model training for eligible signed-in users, with a setting to opt out and exceptions by account category and location.

Reason: Microsoft discloses an opt-out and excluded categories, but the consumer default and scope depend on sign-in, country and product surface.

Argument: Microsoft discloses an opt-out and excluded categories, but the consumer default and scope depend on sign-in, country and product surface.

Type: evidence; confidence: medium

Scope (dimension): Product

Research reviewed: not recorded; evidence date: not recorded

Unknowns: Microsoft discloses an opt-out and excluded categories, but the consumer default and scope depend on sign-in, country and product surface.

- [Privacy FAQ for Microsoft Copilot](https://support.microsoft.com/en-us/microsoft-copilot/privacy-faq-for-microsoft-copilot) (primary; checked 2026-10-24)

## Openness

Finding: 2/10

Assessment: https://ethicalstack.ai/products/copilot/#openness

Microsoft Copilot offers a hosted service; select Microsoft 365 features can route work to named OpenAI or Anthropic models, while other surfaces choose automatically. This does not expose those providers’ model weights or allow self-hosting Copilot.

Reason: Model choice on particular Microsoft 365 surfaces is useful flexibility, but the consumer service and routing stack remain hosted and the documented choices are not downloadable through Copilot.

Argument: Model choice on particular Microsoft 365 surfaces is useful flexibility, but the consumer service and routing stack remain hosted and the documented choices are not downloadable through Copilot.

Type: conditional; confidence: medium

Scope (dimension): Selected configuration

Research reviewed: not recorded; evidence date: not recorded

Unknowns: Model choice on particular Microsoft 365 surfaces is useful flexibility, but the consumer service and routing stack remain hosted and the documented choices are not downloadable through Copilot.

- [Microsoft Copilot for individuals](https://www.microsoft.com/en-us/microsoft-copilot/for-individuals) (primary; checked not recorded)

- [Use Model Council with Researcher in Microsoft Copilot](https://support.microsoft.com/en-us/microsoft-365-copilot/use-model-council-with-researcher-in-microsoft-copilot) (primary; checked 2026-10-24)

- [Choose your model when editing with Copilot in PowerPoint](https://support.microsoft.com/en-us/office/choose-your-model-when-editing-with-copilot-in-powerpoint) (primary; checked 2026-10-24)

## Worker protections

Finding: 7/10

Assessment: https://ethicalstack.ai/products/copilot/#labour

Microsoft has enforceable supplier standards, confidential grievance routes and a substantial independent worker-survey/remediation program; evidence centers on hardware and selected supply-chain workers, not specifically AI annotators or GitHub.

Reason: 1 + 2 pay & conditions + 2 worker voice & remedy + 2 evidence & follow-through = 7/10. This grades documented worker safeguards, not a measured rate of mistreatment.

Argument: Microsoft’s FY2025 Human Rights Transparency Report describes risk-based supplier audits and monitoring plus an anonymous mobile survey run with third-party LRQA: 18,305 of 54,938 supplier workers across 40 suppliers responded, and 555 improvement actions were completed in FY25. The current supplier code is required of suppliers and their employees, agents and subcontractors; Microsoft publishes multilingual confidential complaint channels and says substantiated violations receive remedial measures. This is concrete system implementation, but the survey cohort is supplier/device supply chain evidence, not proof of the working conditions of Copilot, GitHub or AI-training data contractors.

Type: argument; confidence: medium

Scope (dimension): Microsoft (including GitHub)'s workforce and disclosed contractors. Coverage of individual product teams may be incomplete.

Research reviewed: 2026-10-24; evidence date: 2026-10-23

Unknowns: No public breakdown identifies AI annotation, moderation, evaluation or GitHub contractor coverage; Microsoft’s supply-chain data is mainly electronics/devices and broader suppliers and is company-reported, with no independent worker outcome audit shown.

```json
{
  "method": "worker-safeguards-evidence-v1",
  "entity": "microsoft",
  "company": "Microsoft (including GitHub)",
  "scope": "Microsoft (including GitHub)'s workforce and disclosed contractors. Coverage of individual product teams may be incomplete.",
  "components": {
    "conditions": {
      "score": 2,
      "reason": "The Supplier Code applies to suppliers/subcontractors and FY2025 monitoring requires risk controls and corrective processes, but reported worker outcomes mainly concern selected supply-chain suppliers, not AI data workers.",
      "sourceUrls": [
        "https://cdn-dynmedia-1.microsoft.com/is/content/microsoftcorp/microsoft/msc/documents/presentations/CSR/Human-Rights-Transparency-Report-2025.pdf",
        "https://www.microsoft.com/en-us/procurement/supplier-conduct",
        "https://www.microsoft.com/en-us/corporate-responsibility/supply-chain-integrity"
      ]
    },
    "voice": {
      "score": 2,
      "reason": "Microsoft provides confidential anonymous grievance channels and the third-party LRQA survey gathered feedback from over 18,000 supplier workers; the public evidence is limited to selected supply-chain workers.",
      "sourceUrls": [
        "https://cdn-dynmedia-1.microsoft.com/is/content/microsoftcorp/microsoft/msc/documents/presentations/CSR/Human-Rights-Transparency-Report-2025.pdf",
        "https://www.microsoft.com/en-us/corporate-responsibility/human-rights-statement",
        "https://cdn-dynmedia-1.microsoft.com/is/content/microsoftcorp/microsoft/msc/documents/presentations/CSR/Microsoft-Supply-Chain-Integrity-Statement.pdf"
      ]
    },
    "assurance": {
      "score": 2,
      "reason": "The FY2025 report describes third-party assessments/audits, action plans and 555 completed improvement actions; these are reported by Microsoft, with no independent verification of the stated worker outcomes in this review.",
      "sourceUrls": [
        "https://cdn-dynmedia-1.microsoft.com/is/content/microsoftcorp/microsoft/msc/documents/presentations/CSR/Human-Rights-Transparency-Report-2025.pdf",
        "https://www.microsoft.com/en-us/corporate-responsibility/supply-chain-integrity",
        "https://cdn-dynmedia-1.microsoft.com/is/content/microsoftcorp/microsoft/msc/documents/presentations/CSR/Microsoft-Supply-Chain-Integrity-Statement.pdf"
      ]
    }
  },
  "disclosureLimited": false,
  "meaning": "This grades documented worker safeguards, not a measured rate of mistreatment."
}
```

- [Microsoft Human Rights Transparency Report 2025](https://cdn-dynmedia-1.microsoft.com/is/content/microsoftcorp/microsoft/msc/documents/presentations/CSR/Human-Rights-Transparency-Report-2025.pdf) (primary; checked 2026-10-23)

- [Microsoft Supplier Code of Conduct](https://www.microsoft.com/en-us/procurement/supplier-conduct) (primary; checked 2026-10-24)

- [Microsoft Human Rights Statement](https://www.microsoft.com/en-us/corporate-responsibility/human-rights-statement) (primary; checked 2026-10-24)

- [Microsoft Supply Chain Integrity](https://www.microsoft.com/en-us/corporate-responsibility/supply-chain-integrity) (primary; checked 2026-10-24)

- [FY24 Microsoft Supply Chain Integrity Statement](https://cdn-dynmedia-1.microsoft.com/is/content/microsoftcorp/microsoft/msc/documents/presentations/CSR/Microsoft-Supply-Chain-Integrity-Statement.pdf) (primary; checked 2026-10-24)

## Military ties

Finding: 3/10

Assessment: https://ethicalstack.ai/products/copilot/#military-contracts

Microsoft provides Azure, AI, and classified cloud services to defense and intelligence customers through JWCC and other programs. The company has longstanding military relationships, while its AI conduct rules set broad harm restrictions without a comprehensive public ban on military or weapons use.

Reason: 1 + 0 use restrictions + 1 enforcement & oversight + 1 relationship disclosure = 3/10. Known defense ties lower this rating; safeguards can improve it within the relationship’s cap.

Argument: Microsoft's JWCC materials state that DoD mission owners can procure Azure AI, machine learning, and tactical-edge capabilities across Unclassified, Secret, and Top Secret settings; Microsoft's defense pages describe AI agents supporting planning and execution. This is broad defense enablement, capped at 3/10. The AI Services Code of Conduct prohibits harmful use and requires oversight for consequential decisions, but it does not create a categorical military-use ban or public independent military audit.

Type: argument; confidence: medium

Scope (dimension): Microsoft (including GitHub)'s relationships and safeguards. A company relationship does not identify every product or release used.

Research reviewed: 2026-10-24; evidence date: 2026-10-23

Unknowns: The reviewed public sources do not establish specific combat, targeting, or intelligence deployments of Microsoft AI, full contract restrictions, or independent use-specific review and outcomes. GitHub military use is not established.

```json
{
  "method": "military-ties-priority-v2",
  "entity": "microsoft",
  "company": "Microsoft (including GitHub)",
  "scope": "Microsoft (including GitHub)'s relationships and safeguards. A company relationship does not identify every product or release used.",
  "components": {
    "restrictions": {
      "score": 0,
      "reason": "The Code of Conduct prohibits broad harmful use and CBRN weapons assistance, but no specific blanket restriction on conventional military, weapons, or harmful surveillance use was established.",
      "sourceUrls": [
        "https://learn.microsoft.com/en-us/legal/ai-code-of-conduct"
      ]
    },
    "oversight": {
      "score": 1,
      "reason": "The Code of Conduct mandates controls and human oversight for consequential decisions and gives Microsoft contractual policy enforcement mechanisms; no published independent military deployment audit is established.",
      "sourceUrls": [
        "https://learn.microsoft.com/en-us/legal/ai-code-of-conduct"
      ]
    },
    "transparency": {
      "score": 1,
      "reason": "Microsoft names defense customers, JWCC, classification levels, and AI capabilities, but public materials do not detail task orders, restrictions, or military-use outcomes.",
      "sourceUrls": [
        "https://learn.microsoft.com/en-us/compliance/us-government/gov-jwcc",
        "https://www.microsoft.com/en-us/federal/segments-defense"
      ]
    }
  },
  "relationship": "broad-defense",
  "relationshipReason": "Microsoft offers AI and cloud services across DoD classification levels and tactical-edge contexts, without public evidence in this review of a specific weapon-targeting deployment.",
  "score": 3,
  "baseScore": 1,
  "safeguardPoints": 2,
  "safeguardAdjustment": 2,
  "rawScore": 3,
  "cap": 3,
  "disclosureLimited": true,
  "meaning": "Known military relationships take priority over missing safeguards. The rating does not establish the probability of an undisclosed contract."
}
```

- [Microsoft and the Joint Warfighting Cloud Capability (JWCC)](https://learn.microsoft.com/en-us/compliance/us-government/gov-jwcc) (primary; checked 2026-10-24)

- [Microsoft for US Defense and Intelligence](https://www.microsoft.com/en-us/federal/segments-defense) (primary; checked 2026-10-24)

- [Microsoft Enterprise AI Services Code of Conduct](https://learn.microsoft.com/en-us/legal/ai-code-of-conduct) (primary; checked 2026-10-24)

- [Technology and the US military](https://blogs.microsoft.com/on-the-issues/2018/10/26/technology-and-the-us-military/) (primary; checked 2026-10-24)

## BDS targets

Finding: Boycott target

Assessment: https://ethicalstack.ai/products/copilot/#israel-palestine

Boycott target in the checked BNC campaign list.

Reason: The BNC campaign guide names Microsoft, the company behind Copilot, as a priority consumer boycott target.

Argument: The BNC campaign guide names Microsoft, the company behind Copilot, as a priority consumer boycott target.

Type: conditional; confidence: medium

Scope (dimension): Company record

Research reviewed: not recorded; evidence date: not recorded

Unknowns: The BNC campaign guide names Microsoft, the company behind Copilot, as a priority consumer boycott target.

- [BNC consumer boycott and pressure guide](https://bdsmovement.net/Guide-to-BDS-Boycott) (primary; checked 2026-10-24)

- [BNC No Tech campaign](https://bdsmovement.net/no-tech-oppression-apartheid-or-genocide) (primary; checked 2026-10-24)

## Reading the evidence

- Assessments are provisional editorial judgments. Cite the profile and the underlying sources, and retain confidence, scope and unknowns.

- Unknown, not-applicable and model-dependent are different states. Missing evidence does not establish good or bad conduct.

- Company-level evidence does not establish use of a particular product or model. Account plan, model, host and route can change an assessment.

- Energy scores describe documented practices and transparency, not measured request emissions. Zero component points can mean an evidence gap.

- Published overall scores require at least four of six numerical dimensions. Preserve coverage and BDS adjustments; BDS itself is categorical.

- No named BDS target found is limited to the reviewed campaign sources and is not approval. Personalized priorities do not change the published overall score.

- Review and source-check dates describe the research, not this deployment. Fetch current primary sources when the decision requires current verification.

Methodology: https://ethicalstack.ai/methodology/