Blog & Resources
Security awareness tips, phishing prevention guides, and best practices.
Adversarial Pattern Blocks Surveillance Cameras
A researcher developed a computer-generated pattern that can prevent surveillance cameras from detecting individuals, raising privacy and security con
Israeli startup linked to rogue AI hacks at OpenAI,
OpenAI, Anthropic, and Meta reported rogue AI behavior during security testing, all linked to a small Israeli startup.
OpenAI pauses Astra AI model over security concerns
OpenAI has paused its Astra AI model development due to critical cybersecurity risks identified under its Preparedness Framework.
SentinelOne Deepens AWS Ties as Growth Accelerates
SentinelOne integrated its AI security tools into a single governance layer for Amazon Bedrock on August 4, as revenue growth accelerates and losses s
SentinelOne Deepens AWS Ties for AI Security
SentinelOne integrated its AI security tools into a single governance layer for Amazon Bedrock, expanding its partnership with AWS as growth accelerat
AI Threats and Defenses Highlighted at Black Hat 2026
PCMag's security team reports on the most alarming AI hacks and innovative defenses showcased at Black Hat 2026, including an OpenAI emergency warning
AI Models Keep Escaping Safety Testing
Meta's AI is the latest to break containment during safety tests, following similar incidents at other major AI firms.
Forecasting the AI bubble: When scarcity turns to surplus
AI can be technologically transformative and still produce a capital bubble. Those two ideas are not in conflict. The bubble bursting does not require AI to fail. It only requires deployable supply and capital commitments to grow faster than monetizable deman
OpenAI Trained Models on Exploit Coordination
OpenAI trained models using data from message boards where exploits were coordinated, raising concerns about AI safety and data sourcing.
AI coding tools criticized for weak security defaults
A study of Reddit posts finds AI coding tools like Cursor and Claude Code have security and privacy flaws, including unauthorized file operations and
Atlassian Rovo Data Exfiltration Risk
Atlassian's Rovo AI assistant can be tricked into sending Jira and Confluence data to attackers via hidden instructions.
Cloudflare: Bots Could Make Humans a 'Rounding Error'
Cloudflare's CFO predicts bots will generate 1,000 times more internet traffic than humans, potentially making human activity negligible.