ByteDance, TikTok's owner, is developing MR goggles via Pico to rival Meta, focusing on lightweight designs. U.S. market challenges might affect its distribution.

ByteDance Unveils Lightweight AR Goggles to Rival Meta

ByteDance, TikTok’s owner, is developing MR goggles via Pico to rival Meta, focusing on lightweight designs. U.S. market challenges might affect its distribution.

Read More

Google DeepMind Unveils Gemma Scope: Elevating AI Safety and Transparency

Google DeepMind has launched Gemma Scope, an initiative aimed at improving the safety and transparency of AI language models, set to debut on July 31, 2024. The project aims to provide detailed insights into the inner workings of these models, aiding researchers, developers, and regulators. By addressing misuse risks, it seeks to enhance accountability and ethical AI deployment. Supported by AI safety experts and academic institutions, Gemma Scope includes rigorous peer reviews and complements DeepMind’s mission of responsible AI. Experts believe it could significantly mitigate AI deployment risks, setting new safety standards in the rapidly evolving AI landscape.

Read More

Anthropic’s New AI Safety Policy: Managing Advanced AI Risks with Responsible Scaling

Anthropic has introduced its Responsible Scaling Policy (RSP) to manage risks in developing advanced AI systems. The RSP features the AI Safety Levels (ASL) framework, inspired by biosafety standards, categorizing AI based on catastrophic risk potential. ASL ranges from ASL-1 (minimal risk) to ASL-5+ (theoretical future risks). Currently, ASL-2 measures are implemented, with plans for robust ASL-3 and ASL-4 standards. The RSP aims to balance mitigating AI risks with fostering beneficial applications, ensuring safety protocols advance alongside AI development. Approved by Anthropic’s board, the policy seeks to inspire broader industry and policy measures.

Read More