Microsoft has announced a new feature in its Azure OpenAI Service called “Risks & Safety Monitoring” to help businesses detect and prevent abuse of their AI systems. This feature provides real-time information on harmful end-user requests, such as total blocked request count, block rate, and severity distribution by category. It also helps businesses identify users who are consistently making abusive requests and take action accordingly. The monitoring features are currently available for customers in select regions, with no word on when it will be expanded globally. This new feature aims to help businesses maintain control over their AI systems and prevent them from being used against their guidelines.
Source link
Source link: https://www.neowin.net/amp/microsoft-azure-openai-service-now-offers-real-time-detection-of-abusive-requests/
Microsoft Azure OpenAI Service detects abusive requests in real-time. #AIAbuseDetection
