Interactive tools and local models lead October AI launches

Google launched Guided Vision in Gemini Live, providing spoken descriptions and camera-framing guidance developed with blind and low-vision communities. It is not intended for navigation or obstacle detection. Google also released EmbeddingGemma 2 for searching across text, images, audio, video and code, and confirmed its Project Suncatcher prototype satellite entered orbit to test AI computing hardware in space, not establish a commercial orbital data center.

OpenAI expanded ChatGPT access to GPT-6 and Intelligent UI, which combines conversational answers with interactive diagrams, forms and charts. Anthropic launched Claude Haiku 5.5 for repetitive tasks, claiming an average cost reduction of around 75% versus Haiku 4.5, while noting savings depend on workload characteristics.

NVIDIA introduced a DGX Spark 64GB configuration and tools for connecting systems to handle larger workloads. Microsoft made Microsoft Execution Containers generally available on Windows 11 to control agents’ access to files and networks. Microsoft and NVIDIA also announced RTX Spark-based Windows PCs, while local-model integrations and additional Copilot capabilities remained forthcoming.

Anthropic expanded verified access for authorized cybersecurity work without removing safeguards, and updated usage rules covering deception, election interference, surveillance and autonomous physical actions. The policy takes effect on November 12 and requires human oversight for equipment capable of causing injury. OpenAI separately reported banning influence operations attributed to Russia and Iran, cautioning that large volumes of AI-generated content do not establish actual influence over public opinion.