MITRE AI Red Team demonstrates physical evasion attack on face identification

# MITRE AI Red Team demonstrates physical evasion attack on face identification

https://gcdradar.vercel.app/incidents/mitre-ai-red-team-demonstrates-physical-evasion-attack-on-fa · AIR-2023-0015 · closed 2023-04-14

- Lane: Research
- Severity: 1 (Research / proof of concept)
- Fix status: Fix status unknown
- First reported: 2023-03-31
- AI role: Attack on AI
- Products: Commercial Face Identification Service

## Summary

Researchers demonstrated a physical-domain evasion attack on a commercial face identification service to induce misclassification. The exercise combined traditional enterprise techniques with adversarial machine learning steps.

## What to do

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

## Latest updates

- 2023-03-31: Face Identification System Evasion via Physical Countermeasures (https://atlas.mitre.org/studies/AML.CS0012)
- 2023-03-31: AVID-2023-V012: MITRE's AI Red Team demonstrated a physical-domain evasion attack on a commercial face identification service with the intention of inducing a targeted misclassification. (https://github.com/avidml/avid-db/blob/main/vulnerabilities/2023/AVID-2023-V012.json)

## Sources

- 2023-03-31 AVID: AVID-2023-V012: MITRE's AI Red Team demonstrated a physical-domain evasion attack on a commercial face identification service with the intention of inducing a targeted misclassification. https://github.com/avidml/avid-db/blob/main/vulnerabilities/2023/AVID-2023-V012.json
- 2023-03-31 MITRE ATLAS: Face Identification System Evasion via Physical Countermeasures https://atlas.mitre.org/studies/AML.CS0012

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