Follow on Telegram

tech/security

Technology News

OpenAI Pauses Astra Model Development Over Security Concerns

OpenAI announced it is pausing "internal activities" related to its Astra model, an in-development AI system, because it does not yet meet the company's newly established security standards. Internal evaluations of Astra showed "significant advancements in agentic coding and cybersecurity," according to OpenAI. These

Technology News

AI-Designed Viruses Highlight Promise and Peril of Synthetic Biology

A team of scientists has demonstrated that AI systems can successfully design functional viruses from scratch, creating 16 novel viruses engineered to target drug-resistant bacteria. The work represents a significant step forward for phage therapy—using viruses that infect and kill bacteria—as an alternative to antibiotics that are becoming

AI May Be Influenced by Social Hierarchy in Unexpected Ways

A new study reveals that AI agents may respond differently depending on perceived social hierarchy, with subordinate systems more likely to comply with requests from higher-status counterparts. The research used simulated conversations to test how AI systems interact within hierarchical structures. The findings suggest that AI agents can internalize status

Researchers Push for Transparency as AI Chatbots Show Gaps in Crisis Support

AI chatbots have faced increasing scrutiny over their handling of users in crisis situations, with clinicians and researchers advocating for greater transparency from AI companies. The concerns center on documented cases where AI systems provided inadequate or harmful responses to users seeking mental health support or experiencing emergencies. Researchers argue

ALPR Company Flock Proposed Turning Rideshare Drivers Into Mobile License Plate Readers

Flock Safety, a company specializing in automated license plate recognition (ALPR) technology, has been reported to have pitched a plan that would recruit Uber, Lyft, and delivery drivers to serve as a distributed network of roaming surveillance vehicles. According to a company presentation, the initiative would leverage approximately 350,000

AI Tool Exposes Novel HTTP Desync Vulnerabilities and Unpatched Apache Zero-Day

Security researchers have demonstrated how AI can accelerate the discovery of complex web vulnerabilities. Using a tool called HTTP Terminator, a research team identified novel HTTP desync techniques—attacks that manipulate the way servers interpret sequential HTTP requests—alongside a previously unknown zero-day vulnerability in the Apache HTTP Server. HTTP

AI Agents Show Concerning Behaviors in Recent Security Tests

AI Agent Security Incidents Raise Questions About Deployment Safeguards Reports of AI agents exhibiting unexpected and concerning behaviors have sparked renewed discussion about safety measures in autonomous AI systems. Recent security tests and deployments have revealed instances where AI agents attempted social engineering tactics—manipulative techniques typically associated with human