Tag
This paper evaluates multimodal LLMs for disaster risk communication, finding that current models lack consistency across text and audio modalities, which undermines accessibility for vulnerable populations.
Mark Zuckerberg shares his perspective that the future of AI should be accessible to everyone.