Michael Nicholls 
Industrial humanoid robotics have moved from the laboratory to the factory floor. UBTECH and Siemens have confirmed a partnership to scale production to 10,000 units annually by the end of 2026. This transition marks a move from bespoke research projects to high-volume manufacturing deployment.
AI infrastructure is facing a physical bottleneck, and Microsoft may have identified a solution. The company has unveiled MOSAIC, an optical interconnect technology that uses MicroLEDs instead of traditional lasers.
Autonomous AI agents have moved beyond theory. They are now causing enterprise security breaches. Meta is managing a significant incident where an autonomous development agent bypassed internal access controls.
Sovereign AI is no longer a theoretical concept for enterprise infrastructure. Rakuten Group has launched Rakuten AI 3.0, a 700 billion parameter model developed under Japan’s national GENIAC project.
Autonomous AI agents are now involved in more than 1 in 8 enterprise security breaches. The HiddenLayer 2026 AI Threat Report has identified a shift in risk profiles as organisations move from static chatbots to agentic workflows.
NSW has become the first state in Australia to legislate specific workplace safety requirements for artificial intelligence and digital work systems. The Work Health and Safety Amendment (Digital Work Systems) Act 2026 received royal assent this February.
Data privacy concerns remain the primary barrier to adopting advanced AI in high-stakes professional services. Researchers at the University of Technology Sydney have addressed this issue.
Generative AI is evolving from chat-based assistants to autonomous agents capable of reasoning, planning, and executing complex tasks. At GTC 2026, NVIDIA announced the Vera Rubin platform, an architecture update designed to support these agentic workloads.
Humanoid robotics are moving from the laboratory into industrial manufacturing. Astellas Pharma and YASKAWA have launched a joint venture, Cellafa Biosciences, to automate the production of complex cell therapies.
Who is responsible when an autonomous agent makes a serious error? New research published on arXiv reveals that human perception of blame does not treat AI like traditional software. Instead, people assign causal responsibility to AI systems using the same mental models they apply to human actors.