Rubrik has expanded its Project Hourglass initiative with the launch of Rubrik Code Guardian, a security harness built on Anthropic's Claude Mythos 5. The tool scans AI-generated code in isolated air-gapped environments to defend against Mythos-level threats. Partners AHEAD, Trace3, and World Wide Technology joined the alliance. The announcement came at Rubrik's GSI Summit in Goa, and the system is currently in private preview.
Over 86 percent of cybersecurity leaders expect AI agents to outpace organizational security guardrails within a year, highlighting the urgency behind Rubrik's effort. The company also integrates Rubrik Agent Cloud under the Project Hourglass umbrella, signaling a broader push into AI-driven enterprise security.
Anthropic separately launched Glasswing, an open-source security scanner designed to protect critical infrastructure such as power grids and financial systems. The tool is free for businesses and organizations, and Anthropic is expanding its AI security team to support further development. This moves the company deeper into infrastructure-level AI safety.
In weather forecasting, Microsoft's AI model Aurora is helping hurricane forecasters improve storm tracking. Aurora produces forecasts for temperature, rain, and tropical tracks, and some five-day hurricane predictions have outperformed the National Hurricane Center. Forecasters emphasize that AI serves as a support tool rather than a replacement for meteorologists.
European markets saw a rebound in AI stocks alongside falling oil prices. The shift came after President Trump said the US would not attack Iran before the midterm elections. Investors responded positively to the combined effect of lower energy costs and renewed confidence in technology shares.
On the policy front, Sen. Maria Cantwell and Rep. Lori Trahan released AI policy frameworks as Democrats prepare for potential control of Congress. The plans address rogue AI cyberattacks and the upcoming midterm elections, with involvement from Greg Casar, Sen. Bernie Sanders, Reps. Sara Jacobs and Don Beyer. The group aims to establish AI oversight rules if their party wins in November.
Fraud detection efforts are benefiting from AI adoption. A new report found that 59% of companies using three or more AI tools stopped at least 90% of attempted fraud before suffering a loss, compared to only 32% of companies without AI defenses. AI fraud detection outperformed older methods for 77% of companies, and most adopters use real-time risk scoring, automated document verification, and scanning for AI-generated fraud.
In other developments, the AI platform Sidekick earned approval from the US Department of Veterans Affairs to help veterans navigate benefits including the GI Bill, housing, healthcare, and career retraining. Developed by USC researchers drawing on nearly 30 years of virtual-human research, Sidekick connects users to human experts when needed. Meanwhile, AskAnyModel lets users compare answers from up to 50 AI models at once through a web browser, with the Pro Plan priced at $29.97.
Key Takeaways
- Rubrik launched Code Guardian, powered by Anthropic's Claude Mythos 5, as part of Project Hourglass to defend against Mythos-level threats in AI-generated code.
- Over 86% of cybersecurity leaders expect AI agents to outpace organizational security guardrails within a year.
- Anthropic released Glasswing, a free open-source security scanner for protecting critical infrastructure like power grids and financial systems.
- Microsoft's Aurora AI model has outperformed the National Hurricane Center on some five-day hurricane forecasts.
- European markets rose as oil prices fell and AI stocks rebounded following Trump's statement on Iran.
- Democrats including Sen. Cantwell and Rep. Trahan released AI policy frameworks ahead of potential congressional control.
- 59% of companies using three or more AI tools stopped at least 90% of attempted fraud before suffering a loss.
- Sidekick, a conversational AI platform, received VA approval to help veterans navigate benefits like the GI Bill and healthcare.
- AskAnyModel allows users to compare responses from up to 50 AI models for $29.97 through a web browser.
- AI fraud detection outperformed older methods for 77% of companies, with many non-users planning to adopt AI defenses within 12 months.
Artificial Intelligence Gurus
This website uses cookies and similar technologies for essential functions, analytics, personalization, and targeted advertising. Users can learn more about these practices through the provided storage preferences link. The site also allows users to edit images. The content appears across multiple news outlets including Goshen News and Daily Record News.
Artificial Intelligence Gurus
This website uses cookies and similar technologies for essential functions, analytics, personalization, and targeted advertising. Users can learn more about these practices through the provided storage preferences link. The site also allows users to edit images. The content appears across multiple news outlets including Daily Record News.
Rubrik Expands AI Security Partnership With Code Guardian Launch
Rubrik announced the expansion of Project Hourglass to include Rubrik Code Guardian, supported by Anthropic's Claude Mythos 5. Partners AHEAD, Trace3, and World Wide Technology joined the alliance. Code Guardian uses machine learning to detect and prevent cyber threats. The partnership aims to improve AI security and develop more sophisticated AI models.
Rubrik Expands Project Hourglass With New AI Security Tool
Rubrik launched Rubrik Code Guardian, a security harness using Anthropic's Claude Mythos 5, to protect enterprises against Mythos-level threats. The tool scans AI-generated code in isolated air-gapped environments and is part of Project Hourglass, which also includes Rubrik Agent Cloud. The announcement came at Rubrik's GSI Summit in Goa. Over 86 percent of cybersecurity leaders expect AI agents to outpace organizational security guardrails within a year. The system is currently in private preview and not generally available.
Hurricane Forecasters Get Help From AI to Track Storms
AI models are helping hurricane forecasters improve storm tracking by learning from past weather patterns. Microsoft's AI model called Aurora produces forecasts for temperature, rain, and tropical tracks. Some five-day hurricane forecasts from Aurora have outperformed the National Hurricane Center. Forecasters stress that AI is a tool to support meteorologists, not replace them.
Anthropic launches open-source Glasswing security scanner
Anthropic announced a new open-source project called Glasswing to protect critical infrastructure like power grids and financial systems. The tool acts as a security scanner to find vulnerabilities in software and hardware. It is free for businesses and organizations to use. Anthropic is also expanding its AI security team to help develop the project further.
Democrats outline AI policy plans for future Congress
Sen. Maria Cantwell and Rep. Lori Trahan released AI policy frameworks as Democrats prepare for potential control of Congress. The plans address concerns about rogue AI cyberattacks and the upcoming midterm elections. Other Democrats involved include Greg Casar, Sen. Bernie Sanders, Reps. Sara Jacobs and Don Beyer. The group aims to establish AI oversight rules if their party wins in November.
AI platform Sidekick approved to help veterans
A conversational AI platform called Sidekick earned approval from the US Department of Veterans Affairs to support veterans. It helps veterans navigate benefits like the GI Bill, housing, healthcare and career retraining. The tool was developed by researchers at USC including Rizzo, drawing on nearly 30 years of virtual-human research. Sidekick organizes trusted information and connects users to human experts when needed.
AskAnyModel lets you compare 50+ AI models for $30
AskAnyModel AI is a tool that lets users compare answers from up to 50 AI models at once. The Pro Plan costs $29.97 and includes unlimited use of the comparison feature. Users can ask a question and see responses from six models simultaneously. It works across devices through a web browser and covers tasks like coding and image generation.
European markets rise as oil prices fall and AI stocks rebound
European shares rose on Friday due to lower oil prices and improved sentiment around AI stocks. Oil prices fell after President Trump said the US would not attack Iran before the midterm elections. The market boost came as investors responded to both the energy and technology sectors shifting in favorable directions.
AI Fraud Defenses Help Firms Stop Most Attempted Attacks
A new report found that 59% of companies using three or more AI tools stopped at least 90% of attempted fraud before suffering a loss. Only 32% of companies without AI defenses achieved the same result. AI fraud detection outperformed older methods for 77% of companies, and most adopters use real-time risk scoring, automated document verification, and scanning for AI-generated fraud. Many non-users plan to adopt AI defenses within 12 months, while others aim to strengthen supplier verification and payment controls.
Sources
- Artificial Intelligence Gurus
- Artificial Intelligence Gurus
- Rubrik Expands AI Security Partnership With Code Guardian Launch
- Rubrik expands ‘Project Hourglass’ with new AI security tool to counter Mythos-level threats
- Hurricane forecasters get help from AI to track storms
- Anthropic expands AI security plans for critical infrastructure
- Democrats write AI rules for a Congress they don’t control, yet
- AI Platform Earns VA Approval to Support Veterans - USC Viterbi
- This $30 tool lets you compare 50+ AI models without paying for each one
- European Midday Briefing : Oil Prices Ease, AI Stocks Recover
- 59% of Firms With AI Fraud Defenses Stop Losses Before They Start
Comments
Please log in to post a comment.