Trust Engine v2.4
AI validation & integrity
Google-normalized English answers scored for inconsistency, duplicates, velocity, and geo mismatch.
92
/ 100
Overall trust
Excellent integrity
5 active flags / 0 loaded responses
+2 pts vs last month
Trust dimensions
Six-axis scoring per response
Anomaly feed
5 active flags - translated to English before scoring when needed
Duplicate device fingerprint
PM Awas Yojana Lucknow Cluster
high
47
Sample rule output shown until database responses are loaded and scanned.
Speed-running responses (<8s)
Ayushman Bharat Kolkata Metro
medium
132
Sample rule output shown until database responses are loaded and scanned.
Geolocation mismatch
PM Awas Yojana Patna East
high
19
Sample rule output shown until database responses are loaded and scanned.
Voice pattern repetition
Krishi Sinchayee Ludhiana
low
8
Sample rule output shown until database responses are loaded and scanned.
Conflicting answer logic
Digital India Bengaluru Urban
medium
64
Sample rule output shown until database responses are loaded and scanned.
Response validation workspace
All respondent answers
Inspect who answered, where they answered from, when they submitted, and every saved answer.
Loaded responses
0
Unique respondents
0
Completed
0
Avg answers
0
No respondent answers found for this filter.
Select a response to inspect every answer.
Active validation rules
12 rules enabled - auto-tuned weekly
Block IPs seen on more than 50 surveys
Reject responses faster than 8 seconds end-to-end
Flag identical answer strings within a 2km cluster
Normalize non-English answers to English using Google Translate
Cross-check geo-IP with declared state
Detect copy-paste from training data
Identify dialect/script mismatches
Flag enumerator patterns deviating > 2 sigma