Independent AI index
for professionals

Home / Security / Copilot for cybersecurity

Microsoft

How safe is Copilot for cybersecurity?

GDPR profileOur summary of what the provider itself publishes about your data: EU hosting possible, no training on your data for business use, and self-hosting available. Caution is given to a model with a documented issue, such as an investigation by a regulator. How we measure for cybersecurityCaution

Copilot has a concrete GDPR concern. The Claude models within Copilot are officially outside the EU Data Boundary. If you want strictly EU-only processing, exclude the Claude models in the admin panel. Weigh this heavily before deploying it in cybersecurity.

Where does Copilot stand in cybersecurity?

Shared spot 19 of 25 options for cybersecurity (professional tools and general assistants) on GDPR profile. 7 options, like Copilot, have a Caution profile. Each tile is an option we follow for this field; click for the check.

Strong5
Good7
Limited1
Unknown5
Caution7

Published safeguards

✓EU hosting possibleYes
✓No training on your data (business)Yes
–Self-hostable (on-premise)Not known

Microsoft 365 Copilot (business) falls under the EU Data Boundary and does not train on tenant data. Consumer Copilot does train by default, with an opt-out.

How to use Copilot safely

  • Use the business variant and sign a data processing agreement (DPA).
  • Do not enter unnecessary personal data into prompts; anonymise where possible.
  • Enable EU data residency where the provider offers it.
  • Keep a professional in the loop: a security tool sees everything: log files, emails, usernames and network traffic.

Stronger or equally strong in cybersecurity

SoSafeSoSafe SE (Cologne, Germany)Aikido SecurityAikido Security BV (Ghent, BE)Eye SecurityEye Security (The Hague, Netherlands)HadrianHadrian Security B.V. (Amsterdam, Netherlands)SekoiaSekoia (Rennes, France)

Latest developments

19 October 2026Claude models in Copilot Cowork process data outside the EU Data Boundary

Copilot Cowork relies on Anthropic's Claude models for demanding tasks. Because that processing takes place outside the EU Data Boundary, Anthropic is switched off by default in European Microsoft 365 tenants. Anyone who turns it on has data processed outside the EU Data Boundary; at the same time, pay-per-use (Copilot Credits) is coming, in the Netherlands postponed until 1 December 2026. source

18 September 2026Plugin4Shell: GitHub Copilot still unpatched (zero-click RCE via plugin updates)

Research lab Air Security publicly disclosed Plugin4Shell on 18 September 2026: a zero-click RCE in the plugin mechanism of AI coding agents. The agent pins a plugin to a commit hash but does not verify that the retrieved code belongs to that hash; whoever controls or takes over the upstream repository can thus have code executed during the automatic background update, without any action by the user. GitHub Copilot had no patch at the time of publication, although the report was made to all vendors in June 2026. GDPR relevance: plugins run with the developer's rights (source code, cloud credentials, SSH keys, internal repositories, production systems). If you process personal data on or via development machines, this directly affects your security measures under Art. 32 GDPR. Measures: inventory installed plugins and the owners of those repositories, disable automatic plugin updates where possible, and restrict credentials on development machines. Note: this concerns the GitHub Copilot coding agent, not Microsoft 365 Copilot. source

15 April 2026Caveats regarding the EU Data Boundary

The Claude models within Microsoft 365 Copilot fall outside the EU Data Boundary; 'flex routing' can route outside the EU during peak load and can be switched off by administrators. source

Questions

How safe is Copilot for cybersecurity?
Copilot has a concrete GDPR concern. The Claude models within Copilot officially fall outside the EU Data Boundary. If you want to process strictly EU-only, exclude the Claude models in the admin panel. Weigh this heavily before deploying it in cybersecurity. There is a concrete GDPR concern: read the details on the page.
Does Copilot train on my data?
Microsoft 365 Copilot (business) falls under the EU Data Boundary and does not train on tenant data. Consumer Copilot does train by default, with an opt-out.
What should I watch out for with AI in cybersecurity?
A security tool sees everything: log files, emails, usernames and network traffic. That is personal data, so you need a data processing agreement and must know where the data is stored. Also ask whether the AI assistant is trained on your data. WithSecure (Finland) says it processes data for Luminen in AWS in Europe and does not train foundation models on customer data. Sekoia stores data in Germany or elsewhere in the EU/EEA and is ISO 27001 certified. SoSafe (France) can run self-hosted. With Microsoft Security Copilot you choose the geo of your workspace, but sharing data with Microsoft is on by default. For Charlotte AI, ESET LiveAI and Mindgard the hosting location is not public. Never let an agent act on production systems without analyst approval, and record who makes the 24-hour report under the Cybersecurity Act when an incident occurs.

Source documents

learn.microsoft.comlearn.microsoft.com

Everything about Copilot Best AI for cybersecurity

Summarised by the editorial team from provider documentation · not legal advice

Last checked: 15 October 2026 · How we check

Next stepAll 25 options for cybersecurityProfessional tools and general assistants side by side, by GDPR profile, price and quality.

Every Friday: what changed in AI for cybersecurity

New models, prices, oversight and tools, with sources and what it means for you. One email a week, unsubscribe with one click.

View a sample

We only use your address for this newsletter. You confirm first by email.