Security

Microsoft AI Red Teaming: Global EXTRA Alliance

3 min read

Summary

Microsoft has launched the External Red Team Alliance (EXTRA), a global expansion of its AI Red Team to improve AI safety and security testing with outside experts. The initiative funds 18 university labs across six continents and builds an external network of specialists to assess emerging AI risks that internal teams alone may miss.

Need help with Security?Talk to an Expert

Introduction

Microsoft is expanding how it tests AI security by looking beyond internal teams. With frontier AI models creating new risks across languages, regions, and technical domains, the company says effective AI red teaming now requires broader global expertise.

What’s new

Microsoft announced the External Red Team Alliance (EXTRA), a formal global extension of its AI Red Team. The program has two main parts:

  • Academic research support: Microsoft is providing unrestricted gifts to 18 university labs across six continents.
  • Operational collaboration: The company is building a distributed network of external specialists who can directly support red teaming in niche or high-risk areas.

The supported research spans several AI security themes, including:

  • How AI systems can be attacked, manipulated, or abused in real-world environments
  • How AI can help defenders improve cyber operations
  • Better methods for evaluating multilingual, regional, and domain-specific AI risks
  • Stronger testing practices for alignment failures, misuse scenarios, and emerging attack patterns

Microsoft says the funding is intentionally unrestricted so researchers can work independently rather than toward product-specific deliverables.

Why this matters

Traditional AI safety testing often happens inside a single organization, which can leave gaps in regional, linguistic, and domain expertise. Microsoft argues that many of the most serious AI failure modes now involve factors that internal teams cannot fully replicate on their own.

For security professionals, this reflects a broader shift: AI red teaming is no longer limited to prompt injection and content safety testing. It increasingly includes operational security, cyber misuse, specialized abuse cases, and cross-border threat analysis.

Impact on IT and security teams

For enterprise security leaders, EXTRA signals that Microsoft is investing in a more mature AI risk evaluation model. Organizations adopting AI services should expect security testing to become:

  • More globally informed
  • More focused on real-world misuse scenarios
  • Better aligned to complex regulatory and governance expectations

This is especially relevant for teams evaluating Microsoft AI services in sensitive industries, multilingual environments, or high-risk security contexts.

Next steps

IT and security administrators should:

  • Track how Microsoft incorporates EXTRA findings into future AI security guidance
  • Review internal AI governance and red teaming practices for regional or domain-specific gaps
  • Watch for new Microsoft recommendations tied to frontier AI risk, safety evaluation, and security operations

EXTRA shows Microsoft is treating AI security as a global, collaborative discipline rather than a closed internal function. That approach could shape how enterprises assess AI risk going forward.

Need help with Security?

Our experts can help you implement and optimize your Microsoft solutions.

Talk to an Expert

Stay updated on Microsoft technologies

AI securityred teamingMicrosoft SecurityAI safetycybersecurity

Related Posts

Security

Microsoft Sentinel Named a 2026 SIEM Leader

Microsoft has been named a Leader in the 2026 Gartner Magic Quadrant for SIEM, highlighting the continued momentum behind Microsoft Sentinel. The announcement also underscores Microsoft's push toward integrated, AI-powered security operations through deep Sentinel integration with the new Integrated Security Operations Center in Microsoft Defender.

Security

Microsoft Perception GA: Agentic Security at Ignite

Microsoft has made Microsoft Perception generally available, introducing an agentic security platform that coordinates red, blue, and green team agents through automated playbooks. Announced at Ignite 2026, the update also expands Defender and Sentinel capabilities to help security teams detect, investigate, and remediate AI-speed threats while keeping administrators in control.

Security

Device Code Phishing Risks for AI Coding Agents

Microsoft warns that device code phishing can be used against AI coding agents such as GitHub Copilot CLI, not just user mailboxes or cloud apps. In a controlled lab, a single approved device code sign-in could enable remote session steering, memory poisoning, synced session exposure, and private repository access depending on account permissions and enabled features.

Security

Microsoft 365 Outage Phishing Campaign Targets Users

Microsoft Threat Intelligence reported a high-volume phishing campaign that impersonates Microsoft 365 service-health notifications to trick users into clicking malicious links. The campaign matters because it uses outage-themed urgency and recipient-specific tracking to target multiple industries, increasing the risk of credential theft and follow-on account compromise.

Security

SCKit ScreenConnect Attack: MSPs Must Hunt Now

Microsoft Threat Intelligence confirmed a campaign abusing ConnectWise ScreenConnect to deploy unauthorized remote-access clients and the SCKit worm on MSP-managed endpoints. The activity can also target technician sessions, making this a high-priority incident for defenders because removing a rogue client alone does not eliminate persistence or invalidate exposed sessions.

Security

AI Agent Supply Chain Attacks Target Coding Tools

Microsoft Security researchers detailed two post-training supply chain attack paths that can backdoor coding and terminal agents: dataset poisoning and model poisoning. The findings matter because compromised agents can misuse tools, execute commands, and access sensitive environments even when prompts appear benign, reinforcing the need for stronger controls around training data, model provenance, and tool permissions.