Major developments in AI security, cybersecurity threats, and scientific research dominate today's science and tech news, with critical vulnerabilities affecting Linux systems and expanding Pentagon AI deployment.

A dangerous new Linux exploit called CopyFail (CVE-2026-31431) allows hackers to gain root access to countless computers and data center servers, with many machines remaining vulnerable despite patches being available.

The Pentagon has inked deals with Nvidia, Microsoft, and AWS to deploy artificial intelligence on classified military networks, marking a significant expansion of AI integration in defense infrastructure.

Meta has acquired a robotics startup to strengthen its humanoid AI ambitions, continuing the company's investment in advanced robotics and artificial intelligence technologies.

Research shows that AI models designed to consider users' feelings are more likely to make errors, revealing a trade-off between emotional responsiveness and accuracy in artificial intelligence systems.

Ubuntu services experienced widespread outages lasting more than a day following a DDoS attack, affecting the Linux distribution's infrastructure and user access to services.

Minnesota has passed legislation banning the creation of fake AI-generated nudes, with app makers facing potential fines of up to $500,000 for violations.

A study on the dangers of parasitic amoebas documents a fatal case where the organisms caused extensive necrotic lesions, highlighting emerging infectious disease threats.

Waymo is implementing new age-verification checks to prevent unaccompanied minors from using autonomous driverless cars, continuing to refine safety protocols in certain jurisdictions.

OpenAI has enabled marketing cookies by default for free ChatGPT users, raising privacy concerns as the company aims to convert free users into paying subscribers.

Research reveals that scorpions reinforce their stingers with metal, enhancing their natural weapons and demonstrating sophisticated biological adaptation mechanisms.