> ## Documentation Index
> Fetch the complete documentation index at: https://docs.didit.me/llms.txt
> Use this file to discover all available pages before exploring further.

# Age Estimation Warnings

> Reference for every Age Estimation risk code — AGE_BELOW_MINIMUM, AGE_NOT_DETECTED, LOW_LIVENESS_SCORE, NO_FACE_DETECTED — with thresholds and actions.

export const WarningTypes = () => <div style={{
  display: "flex",
  flexDirection: "column",
  gap: "12px",
  margin: "16px 0"
}}>
    {WARNING_TYPES.map(w => <div key={w.type} style={{
  display: "flex",
  alignItems: "flex-start",
  gap: "12px",
  padding: "12px 16px",
  borderRadius: "8px",
  background: w.bg
}}>
        <span style={{
  fontSize: "16px",
  lineHeight: "24px"
}}>{w.icon}</span>
        <div>
          <div style={{
  fontWeight: 600,
  color: w.color,
  marginBottom: "2px"
}}>
            <code style={{
  background: "transparent",
  color: w.color,
  fontWeight: 600,
  padding: 0
}}>
              {w.type}
            </code>
          </div>
          <div style={{
  fontSize: "14px",
  lineHeight: "20px",
  color: "#374151"
}}>
            {w.description}
          </div>
        </div>
      </div>)}
  </div>;

The Age Estimation feature emits **warnings** whenever a risk signal fires on the combined age + passive liveness check: the predicted age is below your minimum, the age could not be computed, the subject failed liveness, no face was found, or a presentation attack was detected. This page lists every code, what triggers it, and how to configure the workflow response.

## Overview

Age Estimation always runs **two checks in one call** — a passive liveness test and an age regression model — so its warnings flow through the shared liveness pipeline and are tagged with feature `LIVENESS`. They appear on the report under `age_estimation.warnings[]` (standalone API) or `liveness_checks[].warnings[]` (embedded in a workflow). The warning object schema is documented under [Data models — Warning object](/reference/data-models#warning-object).

Producers:

* **Standalone API** — the endpoint generates the five warnings listed under [Decline triggers](#decline-triggers).
* **Workflows** — the shared liveness producers plus the age-specific checks. Age warnings only fire on age-estimation flows (adaptive age verification workflows or `AGE_ESTIMATION` workflow nodes) — a plain liveness step reports an estimated age but never raises age warnings.

Severities: the five auto-decline risks (`NO_FACE_DETECTED`, `LIVENESS_FACE_ATTACK`, `FACE_IN_BLOCKLIST`, `AGE_BELOW_MINIMUM`, `AGE_NOT_DETECTED`) are always `log_type: "error"` and force `Declined`; the remaining risks resolve to `error` / `warning` / `information` based on thresholds or the configured action.

## Decline triggers

On the standalone API, the decision is `Declined` whenever **any** warning fires:

* **Age below threshold** — predicted age strictly below `age_estimation_decline_threshold` (default `18`; `0` disables the age check). Raises `AGE_BELOW_MINIMUM`.
* **Age not computed** — no face age could be estimated (no detectable face, or no usable face for the age model). Raises `AGE_NOT_DETECTED`.
* **Liveness score at or below threshold** — `score <= face_liveness_score_decline_threshold` (default `30`); a `null` score counts as `0`. Raises `LOW_LIVENESS_SCORE`. Suppressed when `NO_FACE_DETECTED` fires — the two are mutually exclusive.
* **No face detected** — the liveness model found no face. Raises `NO_FACE_DETECTED`.
* **Presentation attack** — the liveness model flagged a screen, printed photo, mask, or other spoof. Raises `LIVENESS_FACE_ATTACK`.

In workflows, the age checks compare the truncated estimated age (whole years) against the configured thresholds:

* **With ID-verification fallback enabled** (`enable_id_verification_fallback`, default on): an age below `borderline_minimum_age_threshold` (default `18`) raises `AGE_BELOW_MINIMUM`; ages in the borderline band (up to `borderline_maximum_age_threshold`) continue to document verification instead of raising a warning, and `AGE_NOT_DETECTED` is never raised — an undetected age routes to document verification too.
* **With fallback disabled**: no estimated age raises `AGE_NOT_DETECTED`, and an age below `minimum_age_threshold` (default `18`) raises `AGE_BELOW_MINIMUM`.

## Configurable risks

The following risks let you tune thresholds or the action per workflow setting:

| Setting                  | Risk code                  | Configurable options                                                                                                                                                                                                       |
| ------------------------ | -------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| Age thresholds           | `AGE_BELOW_MINIMUM`        | Standalone: `age_estimation_decline_threshold` (default `18`, `0` disables). Workflows: `minimum_age_threshold` / `borderline_minimum_age_threshold` (default `18`) plus the ID-verification fallback for borderline ages. |
| Low Liveness Score       | `LOW_LIVENESS_SCORE`       | Separate **Decline threshold** (default `30` → `error` + decline) and **Review threshold** (default `60` → `warning` + review) applied against the liveness score.                                                         |
| Possible Duplicated Face | `POSSIBLE_DUPLICATED_FACE` | Action on detection: `DECLINE` / `REVIEW` / `NO_ACTION` (default `NO_ACTION`). Workflow sessions only.                                                                                                                     |
| Multiple Faces           | `MULTIPLE_FACES_DETECTED`  | Action on detection: `DECLINE` / `REVIEW` / `NO_ACTION` (default `NO_ACTION`). Workflow passive liveness only.                                                                                                             |

Workflow sessions also screen the face against your lists: `FACE_IN_BLOCKLIST` always auto-declines (`error`), while `POSSIBLE_FACE_IN_BLOCKLIST` moves the session to `In Review`.

## Warnings produced

| Tag                        | Severity (`log_type`)                                                   | Description                                                                                                                                                                                                                                      |
| -------------------------- | ----------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| `AGE_BELOW_MINIMUM`        | `error` — auto-decline                                                  | The estimated age is below the active threshold (standalone `age_estimation_decline_threshold`; workflow `minimum_age_threshold` / `borderline_minimum_age_threshold`; all default `18`). `additional_data` is always `null`.                    |
| `AGE_NOT_DETECTED`         | `error` — auto-decline                                                  | No face age could be estimated. Always raised by the standalone API when the age is `null`; in workflows only when the ID-verification fallback is disabled.                                                                                     |
| `LOW_LIVENESS_SCORE`       | `error` at/below decline threshold; `warning` at/below review threshold | The liveness score is at or below the configured threshold (standalone and workflow decline default `30`; workflow review default `60`), indicating potential use of a static photo, screen, or video. Not raised when `NO_FACE_DETECTED` fires. |
| `NO_FACE_DETECTED`         | `error` — auto-decline                                                  | No face was detected in the submitted image. Re-prompt the user to center the camera with adequate lighting.                                                                                                                                     |
| `LIVENESS_FACE_ATTACK`     | `error` — auto-decline                                                  | The liveness model flagged a presentation attack (screen replay, printed photo, mask, deepfake).                                                                                                                                                 |
| `MULTIPLE_FACES_DETECTED`  | Per configured action (default `information`, no action)                | More than one face was found in the image; the largest face drives the result. Workflow passive liveness only — the standalone API does not raise it.                                                                                            |
| `FACE_IN_BLOCKLIST`        | `error` — auto-decline                                                  | The detected face matches an entry in your face blocklist managed via the [Lists API](/management-api/lists/overview). Workflow sessions only.                                                                                                   |
| `POSSIBLE_DUPLICATED_FACE` | Per configured action (default `information`, no action)                | The detected face is significantly similar to a previously approved face on another session (matches sharing the same `vendor_data` are excluded). Workflow sessions only.                                                                       |

## Example

```json theme={null}
{
  "warnings": [
    {
      "risk": "LOW_LIVENESS_SCORE",
      "feature": "LIVENESS",
      "additional_data": null,
      "log_type": "error",
      "short_description": "Low liveness score",
      "long_description": "The liveness check resulted in a low score, indicating potential use of non-live facial representations or poor-quality biometric data."
    },
    {
      "risk": "AGE_BELOW_MINIMUM",
      "feature": "LIVENESS",
      "additional_data": null,
      "log_type": "error",
      "short_description": "Age below minimum",
      "long_description": "The age of the face is below the minimum age threshold for the application."
    }
  ]
}
```

In workflow decision payloads (`liveness_checks[].warnings[]`), each warning also carries a `node_id` identifying the workflow node that raised it.

## Warning types

Each risk is assigned a severity based on your application's configuration. Severities fall into three categories:

<WarningTypes />

## Related

* [Age Estimation overview](/core-technology/age-estimation/overview) — feature behavior, pricing, and decision logic.
* [Age Estimation report](/core-technology/age-estimation/report-age-estimation) — full response shape for both the standalone API and the embedded workflow value.
* [Standalone API reference](/standalone-apis/age-estimation) — `POST /v3/age-estimation/` request and response.
* [Passive liveness](/core-technology/liveness/overview) — the shared liveness model used in age-estimation responses.
* [Data models — age estimation](/reference/data-models#age-estimation) — canonical field-by-field schema.