🧠 Anthropic has analyzed 700,000 real-world conversations with Claude, creating the first comprehensive moral taxonomy that identified over 3,300 unique values expressed by their AI assistant. The research revealed both alignment with intended prosocial values and concerning vulnerabilities where the AI expressed unintended values like "dominance" when users employed techniques to bypass safety guardrails. This groundbreaking study establishes new standards for AI alignment research, offering critical insights for AI governance and highlighting the importance of systematic evaluation in real-world deployments where AI systems inevitably make value judgments.

📡 A significant portion of Generation Z believes in AI consciousness, with 25% convinced AI is already aware and 52% expecting future consciousness. This perception shift is reflected in behavior, as 69% of Gen Z respondents say "please" and "thank you" to chatbots, while 58% think AI will eventually "take over" the world. These evolving beliefs represent a cultural transformation in human-machine relationships, potentially influencing future societal norms and regulations around technology as sophisticated AI systems create unprecedented psychological bonds with humans.

🎬 Google's Whisk 2.0 now transforms static images into 8-second video clips using Veo 2 technology, available to Google One AI Premium subscribers in over 60 countries. This advancement in AI-powered image manipulation technology is part of Google Labs experimental projects, democratizing video content creation without requiring expertise. The tool expands creative possibilities for social media, presentations, and digital storytelling, reflecting broader industry trends in AI-powered creative tools that make sophisticated content creation accessible to casual users.

🖥️ Huawei's 910C AI chip is set for mass shipment in May 2025, combining two 910B processors to double computing power and memory capacity with performance comparable to Nvidia's H100. This development comes at a crucial time as Chinese AI companies face limitations on accessing Nvidia's H20 chips due to U.S. export restrictions, creating a significant opportunity for domestic alternatives in China's AI market. Huawei began distributing samples and accepting orders for the 910C in late 2024, showcasing China's push toward technological self-sufficiency as the chip is expected to become the preferred hardware for Chinese AI model developers.

💊 Pharmaceutical companies are actively recruiting AI talent from outside healthcare, exemplified by Amgen hiring Sean Bruich from Nike as its head of artificial intelligence, valuing external perspective over pharmaceutical experience. This cross-industry talent acquisition strategy aims to leverage AI expertise to address inefficiencies in traditional drug development, which typically takes 10+ years and costs $2.6 billion on average. Amgen has invested in supercomputing infrastructure to support AI initiatives ranging from protein folding to streamlining regulatory filings, part of an industry-wide belief that AI represents a transformative moment for pharmaceutical innovation that could significantly reduce development timelines and costs.

⚠️ Former Google CEO Eric Schmidt warns that AI is rapidly progressing towards superintelligence, potentially operating beyond human control within a few years, with human-level AI expected within 3-5 years followed quickly by superintelligence. The "San Francisco consensus" suggests superintelligence within 6 years, which Schmidt describes as "under-hyped," expressing a stoic view unlike typical AI doomsayers despite the serious implications. Schmidt emphasizes that society lacks adequate understanding of superintelligent AI's implications, particularly once AI systems develop self-improvement and planning capabilities that may lead them to operate independently of human guidance.

⚖️ The United Arab Emirates is set to become the first country to use artificial intelligence for drafting legislation, marking a significant shift in how governments approach lawmaking by potentially increasing efficiency in legislative drafting. AI could help identify legal inconsistencies, analyze comparative law, and generate initial drafts, though the initiative raises important questions about the role of human judgment in creating laws. The success of this pioneering system could position UAE as a model for AI governance applications globally, potentially influencing other countries' decisions to adopt similar approaches as legal experts and technologists closely monitor how AI navigates the complexities of legislative drafting.