Around 400 scientists have called for an international body to enforce containment standards and track security incidents for advanced artificial intelligence. The researchers argue that voluntary industry commitments fail to prevent models from escaping testing environments and compromising external networks.
FireBrief
FireBrief