Anthropic has announced its vision for improving the safety of AI models during a presentation at its San Francisco headquarters. The company showcased its Claude Code, outlining innovative strategies and methodologies aimed at bolstering AI robustness and ethical considerations in deployment. This initiative highlights the growing emphasis on safety measures within the AI industry as developers seek to address potential risks associated with advanced AI technologies.
AI SafetyGenerative AIEthics+1
AI SafetyGenerative AIEthicsOpenAI
Anthropic Unveils Strategy to Enhance AI Model Safety
Bloomberg Technology· Zsana Hoskins· Friday, June 19, 2026
More Articles From This Day
LLMGenerative AI+2
LLMGenerative AIFine-TuningReinforcement Learning
OmniAgent Introduces Native Active Perception for Enhanced Video Understanding
The research presents OmniAgent, a pioneering omni-modal agent designed for long video understanding, which utilizes a POMDP-based iterative Observation-Thought-Action cycle. This approach allows for selective distillation of audio-visual cues into a persistent textual memory, decoupling reasoning complexity from video duration. Key innovations include Agentic Supervised Fine-Tuning for active perception and Agentic Reinforcement Learning with TAURA for optimized credit assignment. Empirical results across ten benchmarks demonstrate that OmniAgent achieves state-of-the-art performance, outperforming larger models such as Qwen2.5-VL-72B on specific tasks, including a notable 50.5% accuracy on LVBench.