Tag: agentic AI safety
Browse our exclusive articles!
Prevent Image Analysis by MLLMs Using Visual Prompt Injection
Protect your images from multi-modal large language models with ImageProtector, a new method that blocks unauthorized AI image analysis.
DACO: Enhancing Safety in Multimodal Large Language Models
Discover DACO, a novel framework improving safety in multimodal large language models using dictionary-aligned concept control and sparse coding.
3D-VCD: Mitigating Hallucinations in 3D Embodied Agents
Discover 3D-VCD, a novel framework that reduces hallucinations in 3D-LLM embodied agents using visual contrastive decoding for safer AI decisions.
Exploiting Denoising Flaws in Diffusion Language Models
Discover how denoising irreversibility exposes vulnerabilities in diffusion language models and explore strategies to enhance AI safety and robustness.
Claude AI Shines at HumanX Conference 2024 Highlights
Discover how Claude AI by Anthropic dominated the HumanX 2024 conference with innovations in safety, ethics, and human-AI collaboration.