Each safety event in the Safety Inbox includes one or more event labels that identify driver behaviors and environmental conditions. Event labels enable administrators to evaluate events, determine risk, and coach drivers based on consistent detection criteria.
|
Safety Inbox View |
Event Summary View |
The label category determines how the label affects the event and how it is used in reporting and calculating the Safety Score:
-
Behavior Labels: Describe the driver’s actions that triggered or defined the event. Behavior labels determine the event type and contribute to reporting and scoring when they are included in Safety Score Configuration.
-
Context Labels: Identify environmental and situational conditions related to a safety event. Samsara applies context labels depending on the event data available:
-
AI Context Labels: Generated by Samsara AI models for media-based events. These labels describe surrounding conditions—such as weather, visibility, or nearby road users—and help administrators interpret event details. AI Context Labels are informational by default and contribute to reporting. They affect Safety Scores only when context-based weighting is configured in Safety Score Configuration.
-
Emergency Response Context Labels: Applied automatically when configured auxiliary inputs for emergency lights or sirens are active during an emergency response.
-
Non-AI Context Labels: Applied automatically based on device settings, diagnostic data, or other non-visual inputs.
-
Administrators can add, edit, or remove behavior labels and AI context labels from the event details page to ensure each event is accurately classified. Updating a label immediately updates how the event is categorized in the Safety Inbox, in coaching workflows, and in reports.
System-defined labels that reflect conditional states—such as Speeding detections, Privacy Mode or Emergency Response—cannot be edited. You can, however, add manual behavior or context labels to supplement these events as needed.
Each behavior label identifies a specific driver action that a Samsara AI Dash Cam detects or that an administrator assigns during event review. Behavior labels are divided into two types based on how they are applied to events:
-
Automatic Labels: Applied when the dash cam detects a behavior. Automatic labels are included in the Safety Inbox, coaching workflows, and reports.
-
Manual-Only Labels: Assigned by an administrator during event review. Manual-only labels are included in coaching and reporting but do not contribute to the Safety Score unless they are explicitly included in Safety Score Configuration.
Note
The following behavior labels are available only for event review and coaching. These labels cannot be configured to influence scoring or display in Safety Score reports: Near Pedestrian Collision, Unsafe U-turn, Unsafe Left Turn, Unsafe Reversing, and Unsafe Parking
|
Category |
Automatic Labels |
Manual-Only Labels |
|---|---|---|
|
Crash |
|
— |
|
Collision Risk |
|
|
|
Harsh Driving |
|
— |
|
Distracted Driving |
|
— |
|
Traffic Signs and Signals |
|
|
|
Policy Violations |
|
|
|
Speeding |
|
|
Emergency Response context labels identify safety events that occur when Emergency Lights or Emergency Alarms auxiliary inputs are active during an emergency response. The labels are informational and do not affect the Safety Score; however, events can be excluded from scoring if you automatically dismiss Safety Events that occur while these inputs are active.
AI Context Labels identify the environmental and situational conditions detected during a safety event. Samsara AI models assign these labels to media-based events using image and sensor analysis to capture conditions such as road, weather, traffic, and visibility.
By default, AI Context Labels are informational and do not affect the Safety Score. When you configure the Safety Score, they can influence how behavior impact is applied in Safety Score calculations.
Note
AI Context Labels only apply to media-based safety events. Devices with Privacy Mode enabled do not generate AI Context Labels.
The following categories list the AI Context Labels supported in the Safety Inbox.
|
Category |
AI Labels |
|---|---|
|
Vulnerable Road User |
|
|
Weather Conditions |
|
|
Road Conditions |
|
|
Driver Behaviors (Contextual) |
|
|
Traffic |
|
|
Passengers |
|
|
Visibility Conditions |
|

Comments
0 comments
Article is closed for comments.