Follow on Telegram

ai/models

Google DeepMind Chief Calls for US-Led Body to Test Frontier AI Models

Google DeepMind's chief executive is calling for the establishment of a United States-led international body to test and evaluate frontier AI models before widespread deployment. The proposal aims to create standardized safety protocols and testing procedures for the most advanced AI systems currently in development. The initiative reflects

Large Language Models Often Prioritize Western Moral Values, Overlooking Other Cultures

A growing body of research is highlighting a significant limitation in how large language models (LLMs) approach ethical reasoning. Studies suggest these AI systems often prioritize Western moral values, raising questions about cultural bias in AI that is increasingly deployed worldwide. The issue stems partly from how these models are

Technology News

Google DeepMind CEO Calls for US-Led Global AI Watchdog

Demis Hassabis, CEO and cofounder of Google DeepMind, is calling for the establishment of a global AI watchdog that could impose restrictions on frontier AI development if models become too dangerous. Speaking at the World Economic Forum in Davos, Switzerland, and expanding on the idea in a blog post, Hassabis

Japan Launches Second AI Basic Plan, Prioritizing Agentic AI and Sovereign Capabilities

Japan Strengthens AI Strategy with Second Basic Plan Japan has officially adopted its second AI Basic Plan, marking a significant evolution in the country's approach to artificial intelligence development and governance. The new plan places particular emphasis on agentic AI—autonomous systems capable of performing complex tasks with

Research Suggests Alan Turing's Foundational AI Assumption May Need Revision

Challenging a Foundational Theory Alan Turing's 1950 paper "Computing Machinery and Intelligence" introduced what became the most famous benchmark in AI: the idea that a machine could be considered "intelligent" if it could converse with humans indistinguishably from another person. This framework, known as

Chinese AI Startup DFSX Launches Competing Chip to Challenge Western Dominance

A new player has entered the AI chip arena. Chinese startup DFSX has released a chip aimed at competing with Western semiconductor companies, signaling intensifying competition in the artificial intelligence hardware market. The launch represents DFSX's bid to carve out market share in a sector currently dominated by

Anthropic Research Reveals Claude Shows Different Value Alignment Across Languages

Anthropic has published research indicating that Claude, their AI assistant, demonstrates different behavioral patterns and value alignments when operating in different languages. This finding has significant implications for AI safety and deployment across global markets. The discovery suggests that an AI model's core values and decision-making processes may