The Department of Energy is exploring a small AI data center at the Pacific Northwest National Laboratory in Richland, Washington, with a potential 2028 target. Managed by Battelle, the facility would support advanced AI computing for scientific and national security missions. Options include modular buildings and sustainable energy sources like solar or nuclear power. The lab is also part of the DOE's Genesis Mission, which aims to accelerate scientific discovery across all 17 national labs using AI.
OX Security launched VibeSec to defend against vulnerabilities in Anthropic's Model Context Protocol (MCP), which can lead to insecure AI-generated code. The tool works at three layers: analyzing prompts to prevent risky code, blocking malicious changes to configuration files, and scanning existing code for exposure. This addresses a growing need as AI coding tools become more powerful but still struggle with security, helping developers avoid dangerous patterns before they reach production.
Meta launched a program called MCI that records mouse movements, clicks, keystrokes, and screenshots from U.S. employees' work computers to train AI models on real work tasks like navigating dropdown menus and using keyboard shortcuts. Meta can do this because U.S. law allows monitoring on company devices, while European rules would block it. This gives Meta a unique advantage over competitors like Microsoft and Google, who cannot use customer data the same way.
TruGen AI released Clara, an AI Sales Development Representative that can give personalized product demos, qualify leads, and book meetings with human sales teams around the clock. Clara has a face, voice, and vision, and can join video calls on Zoom, Google Meet, and Microsoft Teams. She also communicates over Slack and Teams and sends follow-up emails automatically, working continuously across time zones and languages.
Okta's Shiven Ramji warns that many companies are deploying AI agents without knowing where they are or what they do. These 'shadow agents' can connect to systems and access data without proper oversight. Ramji says CIOs need to answer three key questions: where are the agents, what can they connect to, and what can they do. Without clear answers, companies risk data leaks and uncontrolled automation.
Key Takeaways
- The DOE is considering a small AI data center at Pacific Northwest National Lab, potentially operational by 2028, using sustainable energy sources.
- OX Security launched VibeSec to protect against vulnerabilities in Anthropic's Model Context Protocol, analyzing prompts and blocking malicious config changes.
- Meta's MCI program records employee mouse movements, clicks, keystrokes, and screenshots to train AI on real work tasks, leveraging U.S. monitoring laws.
- Meta's approach gives it an advantage over Microsoft and Google, who cannot use customer data similarly due to legal restrictions.
- TruGen AI released Clara, an AI SDR that gives demos, qualifies leads, and books meetings on Zoom, Google Meet, and Microsoft Teams.
- Clara works across time zones and languages, communicating over Slack and Teams and sending follow-up emails automatically.
- Okta warns of 'shadow agents'—AI agents deployed without oversight, risking data leaks and uncontrolled automation.
- CIOs must answer where agents are, what they can connect to, and what they can do to manage risks.
DOE considers AI data center at Pacific Northwest National Lab
The Department of Energy is exploring the possibility of building a small AI data center at the Pacific Northwest National Laboratory in Richland, Washington, as soon as 2028. The facility would support advanced AI computing for scientific and national security missions. Battelle, which manages the lab, is studying options like modular buildings and sustainable energy sources such as solar or nuclear power. The lab is also part of the DOE's Genesis Mission, which aims to use AI to speed up scientific discovery across all 17 national labs.
OX VibeSec protects AI supply chain from Anthropic MCP risks
OX Security launched VibeSec to defend against vulnerabilities in Anthropic's Model Context Protocol, which can lead to insecure AI-generated code. The tool works at three layers: it analyzes prompts to prevent risky code, blocks malicious changes to configuration files, and scans existing code for exposure. This addresses a growing need as AI coding tools become more powerful but still struggle with security. VibeSec helps developers avoid dangerous patterns before they reach production.
Meta uses employee data to train AI agents on real work tasks
Meta launched a program called MCI that records mouse movements, clicks, keystrokes, and screenshots from U.S. employees' work computers. The goal is to train AI models to perform tasks like navigating dropdown menus and using keyboard shortcuts. Meta can do this because U.S. law allows monitoring on company devices, while European rules would block it. This gives Meta a unique advantage over competitors like Microsoft and Google, who cannot use customer data the same way.
TruGen AI launches Clara AI SDR for automated sales demos
TruGen AI released Clara, an AI Sales Development Representative that can give personalized product demos, qualify leads, and book meetings with human sales teams around the clock. Clara has a face, voice, and vision, and can join video calls on Zoom, Google Meet, and Microsoft Teams. She also communicates over Slack and Teams and sends follow-up emails automatically. The tool works continuously across time zones and languages.
Shadow AI agents create hidden risks for enterprises
Okta's Shiven Ramji warns that many companies are deploying AI agents without knowing where they are or what they do. These 'shadow agents' can connect to systems and access data without proper oversight. Ramji says CIOs need to answer three key questions: where are the agents, what can they connect to, and what can they do. Without clear answers, companies risk data leaks and uncontrolled automation.
Sources
- AI data center with a different mission being considered in Tri-Cities
- Securing the AI Supply Chain: How OX VibeSec Defends Against Anthropic MCP Vulnerability
- Meta Is Turning Its Workforce Into An AI Training Moat
- TruGen AI Launches Clara AI SDR — The AI Teammate That Converts Website Traffic Into Sales Qualified Pipeline
- The Rise of 'Shadow AI Agents' Inside Enterprises
Comments
Please log in to post a comment.