Search
OpenAI Unveils New System to Prevent AI Models from Disseminating Dangerous Content
2025-04-17 /
Read about 0 minute
Author：小编
OpenAI has introduced a novel monitoring system for its AI reasoning models, o3 and o4-mini, designed to intercept prompts associated with biological and chemical threats. This system ensures that the models do not offer guidance for potentially harmful activities. Tailored to comprehend OpenAI's content policies, the system operates seamlessly alongside the models, detecting risk-related keywords and directing the models to refrain from offering dangerous advice.
-
C114 Communication Network -
Communication Home
7 X 24 Track global technological trends
