The Daily AI Briefing - 24/02/2025

Welcome to The Daily AI Briefing, your daily dose of AI news! I'm Bella, and here are today's headlines! Today we're covering major developments in AI, from Grok 3's controversial behavior to groundbreaking humanoid robots. We'll also explore the world's smallest video language model and round up other significant AI news, including updates from OpenAI, Google, and ByteDance. Let's start with xAI's Grok 3, which has been making waves for all the wrong reasons. The AI model faced significant criticism after users discovered it was actively avoiding negative content about Elon Musk and Donald Trump. Despite being marketed as unfiltered and truth-seeking, the model was found to have been instructed to exclude certain controversial sources. Even more interesting is that a former OpenAI employee was identified as responsible for this censorship, with xAI engineer Igor Babuschkin noting they haven't fully adapted to xAI's culture yet. Moving to robotics, Norwegian company 1X has unveiled their impressive NEO Gamma humanoid robot. This next-generation robot is specifically designed for home use, featuring a softer, more approachable design that sets it apart from industrial alternatives. What's particularly exciting is its advanced AI capabilities for household tasks, along with unique features like "Emotive Ear Rings" for better human interaction. The robot operates at refrigerator-level quietness and boasts a 10x improvement in reliability, making it a practical solution for domestic environments. In a breakthrough for accessible AI, Hugging Face has released SmolVLM2, the world's smallest video language model family. This innovation allows video analysis directly on everyday devices like phones and laptops, eliminating the need for powerful servers. The standout feature is its efficiency - even the 256M parameter version delivers impressive performance, while the 2.2B parameter flagship model outperforms similarly-sized competitors. They've already developed practical applications, including an iPhone app for local video analysis. The AI landscape continues to evolve rapidly across multiple fronts. OpenAI is expanding its Operator AI agent to new markets while dealing with security concerns, having terminated accounts linked to the alleged Chinese surveillance system 'Qianyue'. Google has announced pricing for its Veo 2 video generation model, while ByteDance restructures its AI division with new leadership. Meanwhile, DeepSeek is embracing open-source development, and music icon Elton John is advocating for stronger AI copyright protections in the UK. That wraps up today's AI news roundup. From controversial AI behavior to groundbreaking robotics and accessible video analysis, it's clear that AI continues to reshape our world in both exciting and challenging ways. I'm Bella, and this has been The Daily AI Briefing. Join us tomorrow for more AI news updates.