October 23, 2025

analysis  
Amazon, Microsoft, OpenAI Copilot Security & New Tools

The AI landscape is rapidly evolving with new platforms and tools emerging to manage and secure artificial intelligence. Rubrik has launched Agent Cloud, a platform designed to help companies safely deploy and manage AI agents. This system includes features like Agent Monitor for tracking AI agents across services such as Azure, AWS, and Microsoft 365, Agent Govern for real-time policy enforcement, and Agent Remediate, which can undo harmful AI actions. Rubrik CEO Bipul Sinha emphasized the necessity of understanding and controlling AI agents. Concurrently, Noma Security introduced its Agentic Risk Map (ARM), which visualizes the potential impact of autonomous AI agents across enterprise systems, supporting platforms like Microsoft Copilot Studio and AWS Bedrock. Snyk has also entered the fray with Evo, a system for securing AI-native applications through agent orchestration, addressing risks like prompt injection and shadow AI. On the data front, Reltio launched AgentFlow to connect AI agents to unified, governed data, ensuring they operate with accurate, real-time information. In a different vein, OpenAI is exploring new ways for users to interact with online information by introducing ChatGPT Atlas, a web browser designed to integrate ChatGPT's capabilities for summarizing pages and generating insights, though users are advised to proceed with caution. The educational sector is also seeing AI integration, with Gies College of Business using AI-driven avatars and tools to update its curriculum, and the Better Business Bureau offering AI courses for small businesses. Artist Stephanie Dinkins is employing AI in public art projects to help underrepresented groups teach machines about their identities. Meanwhile, discussions continue regarding the economic implications of AI, with Nobel laureate Richard Thaler cautioning that AI profits are not guaranteed, and Microsoft co-founder Bill Gates predicting that only jobs in energy management, biological sciences, and AI programming will remain resilient against automation. Experts also note that realizing AI's full value requires fundamental changes to operating models, not just applying AI to existing processes.

Key Takeaways

  • Rubrik launched Agent Cloud to manage and secure AI agents, featuring Agent Monitor, Agent Govern, and Agent Remediate.
  • Noma Security unveiled Agentic Risk Map (ARM) to visualize AI agent risks across enterprise systems, supporting platforms like Microsoft Copilot Studio and AWS Bedrock.
  • Snyk introduced Evo for securing AI-native applications through agent orchestration, addressing risks like prompt injection.
  • Reltio launched AgentFlow to connect AI agents to unified data, ensuring they operate with accurate, real-time information.
  • OpenAI introduced ChatGPT Atlas, a new web browser integrating ChatGPT's capabilities for enhanced internet navigation and information summarization.
  • Gies College of Business is using AI-driven avatars and tools to update its curriculum, while the BBB offers AI courses for small businesses.
  • Artist Stephanie Dinkins uses AI in public art projects to help underrepresented groups teach machines about their identities.
  • Economist Richard Thaler warns that profits from the current AI boom are not guaranteed.
  • Bill Gates predicts only jobs in energy management, biological sciences, and AI programming will be resilient against AI automation.
  • Achieving AI value requires organizations to change their operating models, not just apply AI to existing processes.

Rubrik Agent Cloud secures enterprise AI with built-in safety features

Rubrik launched its Agent Cloud platform on October 22, 2025, to help companies safely use AI. Built on the Rubrik Platform, it combines data, identity, and application context for security and efficiency. The platform offers Agent Monitor to find and track AI agents across different services like Azure, AWS, and Microsoft 365. Agent Govern helps control agent actions and enforce policies in real-time. Agent Remediate, using Agent Rewind, allows organizations to undo harmful AI actions without data loss or downtime.

Rubrik introduces Agent Cloud for AI agent security and management

Rubrik has launched Agent Cloud to help organizations manage and secure their AI agents, addressing risks like errors and security threats. The platform, built on the Rubrik Platform, offers security, accuracy, and efficiency for AI transformation. It features Agent Monitor to discover and track agents across various platforms like OpenAI and Amazon Bedrock, Agent Govern to control agent behavior and policies, and Agent Remediate to undo unwanted actions. This comprehensive solution aims to provide peace of mind for businesses adopting AI agents.

Noma Security unveils Agentic Risk Map for AI agent security

Noma Security launched its Agentic Risk Map (ARM) on October 22, 2025, as part of its AI security platform. ARM visualizes the potential impact, or blast radius, of autonomous AI agents across enterprise systems. The platform discovers all AI agents, assesses their security posture, and provides runtime protection. It maps agent connections, tool access, and system dependencies to expose risks. Noma's solution aims to give organizations visibility and control over AI agent security, supporting platforms like Microsoft Copilot Studio and AWS Bedrock.

Rubrik Agent Cloud offers secure AI agent deployment

Rubrik has released Agent Cloud, a platform designed for the secure deployment and management of AI agents. It includes Agent Monitor for automatic agent discovery and activity tracking, Agent Govern for real-time policy enforcement, and Agent Remediate with an Agent Rewind feature to undo unwanted AI actions without data loss. CEO Bipul Sinha highlighted the need to understand and control AI agents, noting their potential for both speed and damage. The platform builds on Rubrik's expertise in data, identity, and resilience and is currently in limited early access.

Snyk Evo secures AI-native applications with agent orchestration

Snyk has launched Evo, a system for securing AI-native applications by orchestrating security agents. This platform addresses risks from generative AI and agentic systems, including prompt injection and shadow AI. Evo operates using the OODA loop framework and automates workflows by combining specialized Task Agents. It features a Policy Agent for natural language policy enforcement and various Task Agents for discovery, threat modeling, and testing. Evo is currently in experimental preview, with wider availability planned for early 2026.

OpenAI introduces ChatGPT Atlas web browser

OpenAI has launched ChatGPT Atlas, a new web browser designed to enhance internet navigation using AI. The browser integrates ChatGPT's capabilities to summarize web pages, answer questions about content, and generate insights from online information. This tool aims to boost user productivity and improve how people find and understand information online. ChatGPT Atlas is currently in early development, with future updates and wider access planned.

Caution advised for OpenAI's new ChatGPT Atlas browser

OpenAI, the creator of ChatGPT, has released a new web browser called ChatGPT Atlas. This browser aims to provide a smarter internet browsing experience by integrating AI capabilities. While it promises enhanced navigation, users are advised to use it with caution. The article suggests that this new tool could significantly change how people interact with online information.

Reltio AgentFlow powers enterprise AI with unified data

Reltio launched AgentFlow on October 21, 2025, to help enterprises use AI agents effectively by connecting them to unified and governed data. The platform addresses the challenge of AI initiatives stalling due to siloed or outdated data. AgentFlow connects AI agents to trusted data from the Reltio Data Cloud through a governed execution layer. This ensures agents act on real-time, context-rich information with built-in security and compliance. Companies like Radisson Hotel Group and Eaton Corporation are piloting the platform.

AI security tech showcased at Songdo Expo for national security

The AI Security Expo in Songdo, Incheon, highlighted advanced AI technologies for national security and public safety on October 22, 2025. The event featured AI-powered video analysis systems to identify suspicious activities in real-time, aiming to improve police analysis and response times. It also showcased AI cybersecurity solutions for threat detection and data protection. Organizers aimed to secure 8 billion won in deals, fostering collaboration between tech developers and government agencies.

Gies College uses AI avatars for radical curriculum changes

Gies College of Business at the University of Illinois Urbana-Champaign is revolutionizing its curriculum using AI, including AI-driven avatars for teaching. Launched in spring 2025, the AI in Business course features an avatar of Professor Robert Brunner that delivers content. This approach, along with AI-generated quiz questions and tools like Coursera's AI-powered learning features, helps keep the curriculum relevant. The college also emphasizes industry engagement and continuous improvement to prepare students for the evolving business world.

BBB Learning offers AI courses for small businesses

The Better Business Bureau (BBB) is offering AI courses through BBB Learning powered by MindEdge to help small business owners navigate the age of AI. These courses cover how AI summarizes search engine results and how to harness AI ethically. The BBB also provides lessons on leadership and other business skills. To encourage participation, they are offering discounts for small businesses during October. The goal is to help businesses adapt to technological changes and improve their operations.

Stephanie Dinkins uses AI for art and storytelling

Artist Stephanie Dinkins is using AI in her public project 'If We Don't, Who Will?' in Brooklyn, where visitor stories are transformed into visual art by AI. The project, launched October 22, 2025, aims to help underrepresented groups teach machines about their identities. The accompanying app is multilingual, using ChatGPT for initial translations. Her other project, 'Data Trust,' encodes community narratives into DNA within living plants, merging culture, memory, and environment.

Economist warns AI profits are not guaranteed

Nobel laureate economist Richard Thaler cautions investors that profits from the current AI boom are not guaranteed. While acknowledging the significant hype around AI, he suggests that profitability may be more complex than anticipated. Thaler's perspective, based on behavioral economics, highlights uncertainties in market dynamics, competition, and AI implementation. His warning reminds investors that hype alone does not ensure financial returns, emphasizing the need for thorough analysis of business models and market conditions.

Bill Gates predicts 3 jobs to survive AI automation

Microsoft co-founder Bill Gates believes only three job sectors will remain resilient against widespread AI automation: energy management, biological sciences, and AI programming itself. These fields require complex problem-solving, specialized knowledge, and human creativity that current AI cannot fully replicate. Gates also predicts AI will transform education within five to ten years by enabling personalized learning experiences. He stresses the importance of developing programming skills and digital literacy for future job security.

AI value realization requires operating model changes

Many organizations are not achieving expected value from AI investments because they are applying it to existing processes without changing their operating models. Peter Bendor-Samuel explains that AI requires a fundamental rethinking of how businesses operate, including people, structure, and workflows. Companies successfully leveraging AI have adopted an 'AI-infused operating model' by using a 'future-back' approach, designing their ideal future state and working backward. This requires rigorous, dynamic frameworks to navigate the uncertainties of AI adoption.

Sources

AI security AI agents Rubrik Agent Cloud Noma Security Snyk Evo OpenAI ChatGPT Atlas Reltio AgentFlow AI risk management AI governance AI data security AI automation AI in business AI education AI ethics AI art AI investment AI job market AI operating models AI national security AI cybersecurity