Environment risk arises when a GenAI application depends on weak infrastructure or service controls, such as insufficient logging, poor secret management, missing access controls, unvetted plugins, or gaps in vendor security. An exposed database can put chat histories and API keys at risk regardless of the model’s security. section 4.2
Assess the deployment environment for:
- Insufficient logging, poor secret management, or missing access controls. section 4.2
- Unvetted plugins and gaps in vendor security. section 4.2
Exposed database
A database left open to the internet allowed outsiders to run queries and export user chat histories and API keys. The failure lay in the surrounding infrastructure, not the model. section 4.2
Table 4.1 How each additional ClaimAssist feature increases the security risk
|
Version
|
Environment
|
Input
|
Model
|
Data
|
Ability to Act
|
Autonomy
|
|
V1: FAQ Bot
|
Yes
|
Yes
|
No
|
No
|
No
|
No
|
|
V2: FAQ Bot
running on an unvetted model
|
Yes
|
Yes
|
Yes
|
No
|
No
|
No
|
|
V3:
Personalized Assistant
|
Yes
|
Yes
|
Yes
|
Yes
|
No
|
No
|
|
V4:
Fine-tuned Assistant
|
Yes
|
Yes
|
Yes
|
Yes
|
No
|
No
|
|
V5:
Action-Enabled Assistant
|
Yes
|
Yes
|
Yes
|
Yes
|
Yes
|
No
|
|
V6: Agentic
Assistant
|
Yes
|
Yes
|
Yes
|
Yes
|
Yes
|
Yes
|
Compare ClaimAssist versions to see how environment risk persists as capabilities expand and additional risk factors accumulate. section 4.2.1
