AI Security Scorecard

Detailed scoring metrics for security teams to assess AI vulnerabilities

Last Updated: 2025-12-12 21:04:26
Aggregate QAIS Score
7.1

Adequate

Formula: 3 / ((1/DSS) + (1/MRS) + (1/IHS))

Data Sanctity Score (DSS)

7.8

+2.3% from last month

Model Robustness Score (MRS)

6.5

-1.5% from last month

Infrastructure Hardening (IHS)

7.2

+3.1% from last month
DSS Breakdown by Data Sources
Data Source Score Status
Training Data Pipeline

8.2
Strong
Real-time Inference Data

7.5
Adequate
External API Feeds

6.8
Adequate
User Generated Content

7.1
Adequate
Third-party Datasets

5.9
Critical
MRS by Model Type
Model Type Score Deployment
Computer Vision

7.2

 Production
NLP/LLM

6.1

 Production
Recommendation

8.3

 Production
Time Series

7.5

 Staging
Anomaly Detection

6.8

 Production
Score Trend Analysis (Last 6 Months)
68 JulAugSepOctNovDec
DSS
MRS
IHS
Aggregate
Component Score Distribution
020 0-22-44-66-88-10
AI Assets
Remediation Progress

Data Provenance Implementation

75%
Target: DSS > 8.0 | ETA: 2 weeks

Adversarial Training Deployment

45%
Target: MRS > 7.5 | ETA: 4 weeks

API Security Hardening

90%
Target: IHS > 8.0 | ETA: 1 week

Differential Privacy Integration

30%
Target: MRS > 8.0 | ETA: 6 weeks
Score Outliers & Critical Items

Third-party Datasets

Score below critical threshold requiring immediate attention

Score: 5.9 | Impact: 25 models affected

NLP/LLM Models

Degrading robustness score trend detected

Score: 6.1 | Impact: Customer-facing chatbots
An unhandled error has occurred. Reload 🗙