Allowed severity of self-harm content to and from the Image Question and Answering scenario for the user.
Verified with Microsoft 365/Office 5568.1000 — updated on July 10, 2026
Supported on: At least Windows Server 2016, Windows 10
Path in the GPO console
User Configuration\Administrative Templates\Microsoft Office 2016\AI\Content Safety\Specific\Image Question and Answering Description
This local content safety policy setting adjusts the allowed severity of self-harm content to and from the Image Question and Answering scenario for the user. The allowed severity may range from 0 to 7. 0 means allow no self-harm content. 7 means allow all self-harm content. For more information on content safety severities, please refer to the following documentation: https://learn.microsoft.com/en-us/azure/ai-services/content-safety/concepts/harm-categories?tabs=definitions For local content safety policy, scenario-specific settings are prioritized over general settings and computer settings are prioritized over user settings. If a higher priority policy setting is not configured, then: - If this policy setting is not configured, the allowed severity of self-harm content to and from the Image Question and Answering scenario is determined by lower priority policy settings. - If this policy setting is not configured and lower priority policy settings are also not configured, the allowed severity of self-harm content to and from the Image Question and Answering scenario defaults to 2. For this policy setting, the order of priority is: 1. Allowed severity of self-harm content to and from the Image Question and Answering scenario for the computer. 2. Allowed severity of self-harm content in general for the computer. 3. Allowed severity of self-harm content to and from the Image Question and Answering scenario for the user. 4. Allowed severity of self-harm content in general for the user.
Registry
software\policies\microsoft\office\16.0\common\ai\contentsafety\specific\imagequestionandanswering Export Builder
BETAConfigure the state, scope and options, then generate .reg, PowerShell, Intune and SCCM outputs — or add the setting to a multi-setting collection.
These exports write the registry — this is not a managed GPO. ⓘ
.reg file
Windows Registry Editor Version 5.00
; Exported from gporais.com
; Policy: Allowed severity of self-harm content to and from the Image Question and Answering scenario for the user.
; State: Enabled
; Supported on: At least Windows Server 2016, Windows 10
[HKEY_CURRENT_USER\software\policies\microsoft\office\16.0\common\ai\contentsafety\specific\imagequestionandanswering]
"allowedselfharmcontent"=dword:00000002More formats (PowerShell, Intune, SCCM)
PowerShell
# Exported from gporais.com
# Policy: Allowed severity of self-harm content to and from the Image Question and Answering scenario for the user.
# State: Enabled
# Supported on: At least Windows Server 2016, Windows 10
# Warning: In SYSTEM context (the Intune default), HKCU targets the SYSTEM profile. Run this script using the logged-on credentials.
$path = 'HKCU:\software\policies\microsoft\office\16.0\common\ai\contentsafety\specific\imagequestionandanswering'
New-Item -Path $path -Force | Out-Null
Set-ItemProperty -Path $path -Name 'allowedselfharmcontent' -Value 2 -Type DWordIntune XML
No direct Policy CSP / OMA-URI mapping for this policy. Use the Intune Remediation tab, or ingest the ADMX in Intune.Intune Remediation
# === Detection script ===
# Exported from gporais.com
# Policy: Allowed severity of self-harm content to and from the Image Question and Answering scenario for the user.
# State: Enabled
# Supported on: At least Windows Server 2016, Windows 10
# Warning: In SYSTEM context (the Intune default), HKCU targets the SYSTEM profile. Run this script using the logged-on credentials.
function Test-RegistryValue {
param(
[Parameter(Mandatory = $true)][string]$Path,
[Parameter(Mandatory = $true)][string]$Name,
[object]$Expected,
[ValidateSet('String', 'DWord', 'MultiString')][string]$Kind = 'String',
[switch]$Absent
)
try {
$item = Get-ItemProperty -LiteralPath $Path -Name $Name -ErrorAction Stop
} catch {
return $Absent.IsPresent
}
if ($Absent.IsPresent) { return $false }
$actual = $item.$Name
if ($Kind -eq 'DWord') { return ([int64]$actual) -eq ([int64]$Expected) }
if ($Kind -eq 'MultiString') {
$actualValues = @($actual)
$expectedValues = @($Expected)
if ($actualValues.Count -ne $expectedValues.Count) { return $false }
for ($i = 0; $i -lt $expectedValues.Count; $i++) {
if ([string]$actualValues[$i] -ne [string]$expectedValues[$i]) { return $false }
}
return $true
}
return [string]$actual -eq [string]$Expected
}
$checks = @(
(Test-RegistryValue -Path 'HKCU:\software\policies\microsoft\office\16.0\common\ai\contentsafety\specific\imagequestionandanswering' -Name 'allowedselfharmcontent' -Expected 2 -Kind DWord)
)
if ($checks -notcontains $false) {
Write-Output 'Compliant'
exit 0
}
Write-Output 'Non-compliant'
exit 1
# === Remediation script ===
# Exported from gporais.com
# Policy: Allowed severity of self-harm content to and from the Image Question and Answering scenario for the user.
# State: Enabled
# Supported on: At least Windows Server 2016, Windows 10
# Warning: In SYSTEM context (the Intune default), HKCU targets the SYSTEM profile. Run this script using the logged-on credentials.
$path = 'HKCU:\software\policies\microsoft\office\16.0\common\ai\contentsafety\specific\imagequestionandanswering'
New-Item -Path $path -Force | Out-Null
Set-ItemProperty -Path $path -Name 'allowedselfharmcontent' -Value 2 -Type DWordSCCM scripts
# Exported from gporais.com
# Policy: Allowed severity of self-harm content to and from the Image Question and Answering scenario for the user.
# State: Enabled
# Supported on: At least Windows Server 2016, Windows 10
# SCCM Configuration Item guidance:
# Create a Configuration Item of type "Setting: Script".
# Discovery script: use the Detection script below.
# Remediation script: use the Remediation script below.
# Compliance rule: the Discovery script output equals 'Compliant'.
# === Detection script ===
# Exported from gporais.com
# Policy: Allowed severity of self-harm content to and from the Image Question and Answering scenario for the user.
# State: Enabled
# Supported on: At least Windows Server 2016, Windows 10
# Warning: In SYSTEM context (the Intune default), HKCU targets the SYSTEM profile. Run this script using the logged-on credentials.
function Test-RegistryValue {
param(
[Parameter(Mandatory = $true)][string]$Path,
[Parameter(Mandatory = $true)][string]$Name,
[object]$Expected,
[ValidateSet('String', 'DWord', 'MultiString')][string]$Kind = 'String',
[switch]$Absent
)
try {
$item = Get-ItemProperty -LiteralPath $Path -Name $Name -ErrorAction Stop
} catch {
return $Absent.IsPresent
}
if ($Absent.IsPresent) { return $false }
$actual = $item.$Name
if ($Kind -eq 'DWord') { return ([int64]$actual) -eq ([int64]$Expected) }
if ($Kind -eq 'MultiString') {
$actualValues = @($actual)
$expectedValues = @($Expected)
if ($actualValues.Count -ne $expectedValues.Count) { return $false }
for ($i = 0; $i -lt $expectedValues.Count; $i++) {
if ([string]$actualValues[$i] -ne [string]$expectedValues[$i]) { return $false }
}
return $true
}
return [string]$actual -eq [string]$Expected
}
$checks = @(
(Test-RegistryValue -Path 'HKCU:\software\policies\microsoft\office\16.0\common\ai\contentsafety\specific\imagequestionandanswering' -Name 'allowedselfharmcontent' -Expected 2 -Kind DWord)
)
if ($checks -notcontains $false) {
Write-Output 'Compliant'
exit 0
}
Write-Output 'Non-compliant'
exit 1
# === Remediation script ===
# Exported from gporais.com
# Policy: Allowed severity of self-harm content to and from the Image Question and Answering scenario for the user.
# State: Enabled
# Supported on: At least Windows Server 2016, Windows 10
# Warning: In SYSTEM context (the Intune default), HKCU targets the SYSTEM profile. Run this script using the logged-on credentials.
$path = 'HKCU:\software\policies\microsoft\office\16.0\common\ai\contentsafety\specific\imagequestionandanswering'
New-Item -Path $path -Force | Out-Null
Set-ItemProperty -Path $path -Name 'allowedselfharmcontent' -Value 2 -Type DWordBuilding a multi-setting collection? Add this setting and generate combined .reg / PowerShell / GPO scripts.