gcdRadar
archive/Research/AIR-2020-0028
closedclosed 6 Nov 2020Research1 Research / proof of conceptAI

Microsoft Edge AI Product Evasion Red Team Exercise

Edge AI Product

The Azure Red Team executed an evasion exercise against a Microsoft edge AI product by manipulating target images. The test successfully forced machine learning misclassifications.

first report
23 Oct 2020
last update
23 Oct 2020

Timeline · 1

  1. 23 Oct 2020Microsoft Edge AI Evasion

Sources · 1

Research and catalogues

  1. 2020-10-23
    Microsoft Edge AI Evasion
    MITRE ATLAS · atlas.mitre.org/studies/AML.CS0011

Titles and links belong to their publishers. The summary above is written by gcdRadar in its own words; data is CC BY-SA 4.0.

Classification

lane
Research
AI role
Attack on AI
categories
Adversarial evasion
severity
1 · Research / proof of concept
targets
Microsoft, Enterprise deployment
owasp llm top 10
no entry for these categories
mitre atlas
AML.T0015 Evade AI ModelAML.T0043 Craft Adversarial Data
tags
red-teamedge-aiadversarial-evasion
dates
happened Feb 2020 (approx.) · first reported 23 Oct 2020
origin
MITRE ATLAS
Derived from MITRE ATLAS (Apache 2.0, © MITRE).
record
confidence 0.95 · created 2026-10-18 06:55Z · updated 2026-10-18 06:55Z

Lane, AI role, categories and severity come from the classifier; the OWASP and ATLAS rows are derived from the categories. How these are assigned

Copy a brief

This record as a short Markdown brief, for a ticket, a chat or a coding agent.

the brief as text
# Microsoft Edge AI Product Evasion Red Team Exercise

https://gcdradar.vercel.app/incidents/microsoft-edge-ai-product-evasion-red-team-exercise · AIR-2020-0028 · closed 2020-11-06

- Lane: Research
- Severity: 1 (Research / proof of concept)
- Fix status: Fix status unknown
- First reported: 2020-10-23 (happened 2020-02)
- AI role: Attack on AI
- Products: Edge AI Product

## Summary

The Azure Red Team executed an evasion exercise against a Microsoft edge AI product by manipulating target images. The test successfully forced machine learning misclassifications.

## What to do

No action recorded; read the sources below and the vendor's advisory.

## Latest updates

- 2020-10-23: Microsoft Edge AI Evasion (https://atlas.mitre.org/studies/AML.CS0011)

## Sources

- 2020-10-23 MITRE ATLAS: Microsoft Edge AI Evasion https://atlas.mitre.org/studies/AML.CS0011

Verify against the vendor's advisory before changing production. Data CC BY-SA 4.0, https://gcdradar.vercel.app.

Related

  1. closed
    Azure Microsoft Internal Azure AI Service Red Team Exercise

    Research · since Jan 2020 (approx.)

    AI
  2. closed
    Google Translate Model Extraction and Adversarial Attacks on Translation APIs

    Research · since 30 Apr 2020

    AI
  3. closed
    Commercial Face Identification Service Face Identification System Evasion via Physical Countermeasures

    Research · since 1 Jan 2020

    AI
  4. closed3.7
    ProofPoint Email Protection ProofPoint Email Protection Evasion via Surrogate Model

    Research · since 9 Sep 2019 · updated 31 Mar 2023 · CVE-2019-20634

    AI
  5. closed
    DGA detector Botnet DGA Detection Evasion Using Domain Mutation

    Research · since 1 Jan 2020

    AI
  6. closed
    Palo Alto Networks malware detection system Evasion of Deep Learning Detector for Malware C&C Traffic

    Research · since 1 Jan 2020

    AI