Microsoft warns that Copilot Actions in Windows, now in beta and off by default, can infect devices and pilfer data, prompting concern from security researchers (Dan Goodin/Ars Technica)

TechmemeThursday, November 20, 2025 at 7:10:02 AM
Microsoft warns that Copilot Actions in Windows, now in beta and off by default, can infect devices and pilfer data, prompting concern from security researchers (Dan Goodin/Ars Technica)
  • Microsoft has warned that its Copilot Actions feature, currently in beta and off by default, poses risks of device infection and data theft, prompting concerns from security experts.
  • This warning highlights the challenges Microsoft faces as it integrates AI features into Windows, balancing innovation with user security.
  • As Microsoft aims to create an 'agentic OS' for autonomous AI agents, the potential security vulnerabilities associated with these technologies have sparked significant debate about the safety of AI integration in everyday computing.
— via World Pulse Now AI Editorial System

Was this article worth reading? Share it

Recommended Readings
Zendesk Unveils AI-Powered Employee Service with New Microsoft Integrations
PositiveArtificial Intelligence
Zendesk has announced new integrations that incorporate its AI-driven support services directly into Microsoft 365. This collaboration aims to enhance efficiency, security, and control for IT, HR, and Finance teams, marking a significant step in Zendesk's ongoing partnership with Microsoft.
LTIMindtree Expands Partnership with Microsoft to Accelerate Microsoft Azure Adoption, Drive AI-Powered Transformation
PositiveArtificial Intelligence
LTIMindtree has expanded its partnership with Microsoft to enhance the adoption of Microsoft Azure and drive AI-powered transformation. This collaboration aims to foster enterprise growth through advanced security and trusted governance offerings, positioning both companies to leverage AI technologies effectively.
Integrating Machine Learning into .NET Healthcare Apps
PositiveArtificial Intelligence
Healthcare software is rapidly evolving, with a significant shift towards integrating machine learning (ML) into systems. This integration allows healthcare teams to derive insights from vast amounts of data generated by hospitals, labs, and connected devices, enhancing patient care and decision-making. Developers using .NET can leverage existing applications to incorporate ML, making systems smarter without starting from scratch.
“We’re in an LLM bubble,” Hugging Face CEO says—but not an AI one
NeutralArtificial Intelligence
Hugging Face CEO has stated that the current landscape of large language models (LLMs) resembles a bubble, though he distinguishes it from a broader AI bubble. He emphasizes that the risks associated with AI investments, particularly in manufacturing and other sectors, remain ambiguous.
Azure Developer CLI Adds AI Agent Extension for Foundry Deployments
PositiveArtificial Intelligence
Microsoft has launched a new azd AI agent extension that enables developers to build and deploy Microsoft Foundry agents directly from their local development environments. This extension aims to streamline the development process, enhancing productivity for developers working with Foundry.
Azure Cosmos DB Adds New AI Search and Agentic Capabilities at Ignite
PositiveArtificial Intelligence
At Ignite 2025, Microsoft announced enhancements to Azure Cosmos DB, introducing new AI capabilities including vector and full-text search features, Semantic Reranking, and an MCP Toolkit for agentic applications. These updates aim to improve the efficiency and effectiveness of data management and retrieval.
Critics scoff after Microsoft warns AI feature can infect machines and pilfer data
NegativeArtificial Intelligence
Microsoft has issued a warning regarding its AI feature, Copilot Actions, which is currently disabled by default in Windows. Critics have raised concerns about the potential risks of this feature, including the possibility of infecting machines and stealing sensitive data. The company has yet to clarify how long this default setting will remain in place.
OpenAI unveils GPT-5.1-Codex-Max, saying it is "significantly better" at "long-horizon reasoning" and is the first model it has trained for Windows environments (David Gewirtz/ZDNET)
PositiveArtificial Intelligence
OpenAI has introduced GPT-5.1-Codex-Max, claiming it is significantly better at long-horizon reasoning and is the first model specifically trained for Windows environments. This new model aims to enhance coding efficiency and tackle complex engineering tasks that may take up to twenty-four hours to complete.