OpenAI launches new model afternew model after US government freeze
OpenAI launched a new AI model following a U.S. government freeze on releases, as the White House divided over how to handle increasingly powerful models.
The government's crackdown aims to prevent risky models from reaching adversaries but may inadvertently push open-source development underground. This shift could reshape global AI policy by forcing nations like China to restrict their own open models.
Western Media
Western media and tech experts argue that the U.S. freeze on AI releases is a necessary security measure to prevent risky models from reaching adversaries, while warning it may inadvertently push open-source development underground.
Chinese Perspective
Chinese observers note that the U.S. crackdown could eventually force China to restrict its own open AI models if they become too powerful, creating a global shift toward closed AI systems.
- The U.S. AI freeze targets models with over 100 billion parameters, a threshold set by the 2026 National AI Security Act.
- OpenAI's new model uses a hybrid architecture that blends transformer and recurrent layers for improved efficiency.
- China's AI regulatory body has pledged to review open models every six months to assess potential security risks.
