Tag: model safety
2 articles
OpenAI Halts Astra Model Tests Over Advanced Cyber Capabilities
OpenAI is hitting the pause button on internal Astra activities that don't meet its new, stricter security control requirements, following concerns over the model's advanced cyber capabilities. The company is implementing enhanced safeguards, including isolated testing environments and encryption, to ensure responsible development.
Anthropic Bolsters AI Model with Enhanced Reasoning, Security Features
Meet Sonnet 5, Anthropic's latest AI model that's setting a new standard for safety and reliability, outperforming its predecessor with a lower rate of undesirable behaviors and enhanced defenses against malicious requests. This cutting-edge model is designed to be more agentic, accurate, and secure, making it a game-changer for users.