Rising use of AI to scale cyberattacks and a parallel push for AI-driven defense tools. Frontier models expose infrastructure vulnerabilities, prompting open-source patching initiatives and government warnings.
The cybersecurity industry ballooned to $200 billion by selling near-real-time risk detection, yet woefully underinvests in fixing actual vulnerabilities. With global spending projected to exceed half a trillion, the imbalance between finding and fix...
A new application-security service leverages OpenAI models to find and verify software vulnerabilities faster. While the partnership pushes automation in enterprise defense, it raises questions about data sovereignty and control over scanning pipelin...
The Five Eyes intelligence alliance warns that the next generation of AI will supercharge offensive hacking. For organizations running LLMs on-premise, the window to prepare is closing fast.
The partnership aims to bring cutting-edge models to enterprise cyber defense, a domain where data sovereignty and on-premise control are becoming decisive factors.
An internal NSA red-team test reveals that Anthropic’s Mythos AI breached nearly all classified systems within hours. The leaked report explains the U.S. government’s sudden ban on flagship models and reignites the debate on LLM security in on-premis...
OpenAI unveils two security tools, Codex Security and GPT-5.5-Cyber, designed to find and patch vulnerabilities at scale. The lack of deployment details raises sovereignty concerns, prompting organizations to weigh cloud convenience against on-premis...
OpenAI launches Patch the Planet, a Daybreak initiative combining AI and expert review to help open-source maintainers find, validate, and fix security vulnerabilities. A promising step that raises questions for those wanting to replicate the approac...
Two recent campaigns show how open-source code and AI tools, pillars for developers, have become the main intrusion vector. No need to break systems anymore: attackers exploit the trust in components we use daily.
Researchers have uncovered "FortiBleed," a vast cache of stolen credentials for Fortinet firewalls. The dataset contains plaintext usernames, emails, and passwords for nearly 74,000 FortiGate devices across 194 countries, affecting over 21,000 domain...
Dream, an Israeli company specializing in AI and cybersecurity, has completed a $260 million funding round, raising its valuation to $3 billion. This increase, nearly triple its value from 16 months ago, highlights the growing demand and strategic im...
Pentesting, a fundamental cybersecurity practice, is no longer a simple choice between acquiring a tool or hiring an external consultant. Organizations, especially those adopting on-premise strategies, face increasing challenges in balancing investme...
