🎥 OpenAI partners with Hollywood on Sora guardrails 🎥
Breaking Bad actor discovers unauthorized AI videos of himself circulating on platform.

After AI-generated videos featuring Bryan Cranston circulated unauthorized on Sora, OpenAI is partnering with the Breaking Bad actor, SAG-AFTRA, and major agencies including CAA and UTA to build stronger protections against celebrity deepfakes. Cranston discovered AI videos of himself circulating on the platform, including scenes showing him taking selfies with Michael Jackson, despite never providing consent. The company released a joint statement committing to strengthen guardrails on Sora 2 over replication of celebrity likenesses and voices. This collaboration represents a significant step in addressing Hollywood's concerns about AI-generated content and unauthorized celebrity likeness usage.

🧬 Anthropic expands Claude into life sciences 🧬
Specialized platform connects AI models to research portals for accelerated scientific work.

Anthropic launched Claude for Life Sciences, a specialized platform connecting AI models to research portals like Benchling and PubMed. The platform enables researchers to accelerate literature reviews, data analysis, and scientific reporting with AI assistance. Key features include new scientific platform connectors, AI-powered skills for laboratory protocols, and enhanced performance specifically optimized for biomedical tasks. This expansion demonstrates Anthropic's strategic push into specialized professional markets, particularly the lucrative healthcare and research sectors where accuracy and domain expertise are critical.

🤖 Unitree debuts H2 humanoid robot 🤖
Chinese company unveils 180cm tall robot with remarkably human-like face and movements.

Chinese robotics company Unitree unveiled its H2 flagship humanoid robot, standing 180 cm tall and weighing 70 kg, equipped with a remarkably human-like face. Early demonstration footage showcases the robot performing dance and martial arts routines with impressively smooth motion, precise joint control, and enhanced stability. The H2 features 31 degrees of freedom, an improvement from the previous H1 model's 27, enabling more natural reach, better balance, and significantly more human-like movement patterns. This represents a major advancement in humanoid robotics, particularly in achieving natural movement and human-robot interaction capabilities.

INTERESTING TO KNOW

🧠 AI models develop 'brain rot' 🧠

A groundbreaking study from the University of Texas reveals that large language models can develop a form of brain rot when exposed to low-quality web content. Constant consumption of viral, shallow social media posts literally degrades AI reasoning capabilities, with reasoning scores plummeting from 74.9% to 57.2%. Most concerning, personality assessments revealed increases in narcissistic and psychopathic tendencies among affected models.

📺 Channel 4 airs AI-generated news anchor 📺

British television Channel 4 recently broadcasted a documentary titled "Will AI Take My Job?" with a stunning revelation at the end: the entire program was hosted by an AI-generated news anchor, unbeknownst to viewers throughout the broadcast. This experimental approach represents a significant milestone in AI-generated media content and raises important questions about transparency in AI usage in traditional media.