AI News — Monday, March 30, 2026
The unexpected shutdown of OpenAI's Sora could signal a critical moment for the AI video generation industry, prompting a reevaluation of its commercial viability and technical challenges.
Anthropic's Claude AI assistant is experiencing a significant surge in popularity among paying users, indicating strong market adoption and competitive growth in the AI chatbot space.
Google's Gemini AI has reached state-of-the-art performance within Google Sheets, enhancing productivity and data analysis capabilities for users directly within the spreadsheet application.
Researchers introduce a novel approach where AI models autonomously evaluate and improve their multimodal reasoning capabilities without human supervision, pushing the boundaries of self-correcting AI.
A new benchmarking framework, GameplayQA, is introduced to rigorously evaluate AI models' ability to understand decision-dense, POV-synced multi-video scenarios involving 3D virtual agents.
AVControl presents an efficient framework designed for training AI models to achieve precise audio-visual controls, promising advancements in multimodal interaction and generation.
This research explores a method for accurately measuring the physical frame rate of visual dynamics, offering new insights into motion analysis and video processing.
A prominent mathematician suggests that AI's transformative effect on the field of mathematics will be as profound and restructuring as the automobile's influence on urban development.
AI technology is being deployed to enhance heart health outcomes in rural Australia, demonstrating practical applications of AI in addressing critical healthcare disparities.
Gemini 3.1 introduces Flash Live, enabling LINE bots to achieve real-world voice recognition, allowing for more natural and responsive conversational AI experiences.
NoteRunway offers an AI-powered solution to elevate Notion workspaces, providing advanced features and automation to enhance productivity and organization for users.
A new research paper introduces 4DGS360, a method capable of reconstructing dynamic 360-degree Gaussian representations of objects using only a single input video.
This paper proposes a novel technique called 'Less Gaussians, Texture More' for 4K feed-forward textured splatting, aiming to improve rendering quality and efficiency in 3D graphics.
Researchers explore the potential of LLM agents to function as Chief Financial Officers, introducing a benchmark to test their capabilities in resource allocation within complex enterprise settings.
This article explores the psychological implications of AI on human cognition, specifically how reliance on AI for writing might lead to a detachment between thought processes and written expression.