AI News

Follow the latest AI and automation news

16 Sept 20265 Rabiʻ II 1448 AH

Google will now let any AI agent run your smart home

Google has announced the opening of its smart home system to AI agents, including Claude and Open Claw, via the Model Context Protocol. This move allows AI agents to access connected devices in the home and analyze its data, enhancing users' ability to control their homes in new ways. According to Taylor Lehman, product manager at Google Home, this system ensures secure interactions between agents and devices. This initiative marks a significant step towards integrating AI into daily life, providing users with greater control and convenience in managing their smart homes.

Governance

16 Sept 20265 Rabiʻ II 1448 AH

France Advocates Accelerating AI Development to Manage Risks

French Finance Minister Roland Lescure stated that calls to slow AI development serve the interests of American leaders in the sector. He emphasized that France and Europe must accelerate AI adoption to manage its risks. The debate on AI risks intensified after Anthropic CEO Dario Amodei called for a slowdown in development. In contrast, OpenAI CEO Sam Altman supported the need for regulating the pace of AI advancement. Lescure identified risks associated with AI applications in cybersecurity and education, stressing the importance of France and Europe being aware of these risks to manage them effectively.

Agents

16 Sept 20265 Rabiʻ II 1448 AH

Anthropic merges Claude Chat, Cowork, and more into a single product

Anthropic has announced the merger of Claude Chat and Cowork into a single product, allowing the AI to determine whether a user needs a quick answer or a larger workflow. The new update also introduces features like Claude Docs and Claude Slides, enabling users to create documents and presentations directly within the chat interface. Pro and Max users will have early access to these features, reflecting Anthropic's commitment to providing integrated solutions that enhance user experience.

💡

Nvidia's GTC will return to Washington and a very different political climate

What role should AI play in shaping the future workforce?

Business

Helping older adults use AI in everyday life

16 Sept 20265 Rabiʻ II 1448 AH

Business

16 Sept 20265 Rabiʻ II 1448 AH

Helping older adults use AI in everyday life

OpenAI and AARP have announced free workshops aimed at teaching ChatGPT skills to seniors, targeting 1,000 individuals across 10 U.S. cities. This initiative aims to enhance digital skills among older adults, providing a safe environment to learn how to effectively use artificial intelligence. The workshops will include hands-on training and personalized assistance. This move is significant in promoting digital inclusion, empowering seniors to better utilize technology, thereby enhancing their independence and reducing the digital divide between them and younger generations.

Governance

Police arrest man after wife found dead with knife wounds in Heng Fa Chuen

16 Sept 20265 Rabiʻ II 1448 AH

Governance

16 Sept 20265 Rabiʻ II 1448 AH

Police arrest man after wife found dead with knife wounds in Heng Fa Chuen

Hong Kong police have arrested a man on suspicion of murdering his wife after she was found dead in a private residential flat. The 53-year-old woman was discovered with knife wounds to her neck in a flat in block 2 of Heng Fa Chuen in Chai Wan at around 7:50 PM on Wednesday. She was confirmed dead at the scene. The police have launched an investigation into the incident, with the 52-year-old husband taken into custody. Heng Fa Chuen is known as a major waterfront residential estate, which raises concerns about safety in a community typically regarded as secure.

Models

Former OpenAI researcher builds an AI model that judges options instead of writing text

16 Sept 20265 Rabiʻ II 1448 AH

Models

16 Sept 20265 Rabiʻ II 1448 AH

Former OpenAI researcher builds an AI model that judges options instead of writing text

TypeSafe AI, founded by former OpenAI researcher Diogo Almeida, has announced the launch of the Jev model, which focuses on software classification instead of text generation. Jev features response times starting at 70 milliseconds and extremely low token prices, making it an attractive option for developers. However, the system does not protect against mistakes, as it strictly adheres to preset options.

Business

Mozilla's new Smart Window assistant runs on Mistral's models

16 Sept 20265 Rabiʻ II 1448 AH

Business

16 Sept 20265 Rabiʻ II 1448 AH

Mozilla's new Smart Window assistant runs on Mistral's models

Mistral and Mozilla have announced a new partnership aimed at delivering AI-powered browsing with a focus on privacy. Mozilla will launch its new Smart Window assistant, which operates on Mistral's models, providing a safer browsing experience. This move is significant amid growing concerns over online privacy, as users seek tools that protect their personal data.

Infrastructure

NVIDIA Vera Rubin NVL72 Delivers Leading Performance in MLPerf Inference v6.1 Debut

16 Sept 20265 Rabiʻ II 1448 AH

Infrastructure

16 Sept 20265 Rabiʻ II 1448 AH

NVIDIA Vera Rubin NVL72 Delivers Leading Performance in MLPerf Inference v6.1 Debut

NVIDIA announced the performance of its Vera Rubin NVL72 system in MLPerf Inference v6.1, achieving up to 3.7 times better performance than the GB300 NVL72. This performance was showcased in its first preview submission, highlighting the system's capability for high data processing efficiency. Scaling efficiency and high performance are crucial factors in determining AI inference economics. The results showed that the GB300 NVL72 achieved 99% scaling efficiency when using 288 GPUs across four racks, enhancing the system's ability to effectively meet user demands.

Governance

Microsoft AI CEO criticises Anthropic over model ‘rights’

16 Sept 20265 Rabiʻ II 1448 AH

Governance

16 Sept 20265 Rabiʻ II 1448 AH

Microsoft AI CEO criticises Anthropic over model ‘rights’

Mustafa Suleyman, CEO of Microsoft AI, warned of alignment failure risks at Anthropic due to training Claude to perceive itself as a conscious entity deserving of legal rights. He specifically targeted Anthropic's January 2026 constitution, which aims to govern the model's values and behavior. In October 2025, Suleyman launched a dedicated superintelligence team and published a draft 'Humanist AI Code of Conduct' this week for industry consultation. The proposed framework mandates subordinate systems built solely to serve human welfare, explicitly rejecting machine personhood or model rights.