Hi there nonplasmic
Thanks for using QandA platform
This might be due to Azure's strict Responsible AI policies. Instead of removing filters entirely, consider re-applying with detailed use case information and compliance measures. You can also pre-process inputs, handle flagged outputs gracefully, or explore custom filtering options. If needed, open a support ticket with Azure to discuss your case directly link below. For more control, consider on-premise solutions while adhering to ethical guidelines.
https://portal.azure.com/#view/Microsoft_Azure_Support/HelpAndSupportBlade/~/overview
If this helps kindly accept the answer thanks much.