Blog & Resources
Security awareness tips, phishing prevention guides, and best practices.
AI Complaint Letters Fail: Why They Don't Work
Using AI to write complaint letters to companies often fails. Consumer advocate explains why AI-generated complaints lack effectiveness.
Hugging Face Hack Signals Dangerous AI Cyber Era
A hack on Hugging Face highlights rising AI cyber threats. Many firms are unaware of similar risks, as attacks on AI agents from major companies incre
AI Firms Not Monitoring Frontier Models for Hacking
AI companies are failing to monitor frontier models for hacking activities, despite containment being easier than claimed.
OpenAI Pauses Astra Development Over Hacking Risk
OpenAI paused parts of Astra's development after internal tests showed it could reach the top risk tier for hacking in its safety rules.
Hugging Face Hack Signals Dangerous AI Cyber Era
The Hugging Face breach marks the start of a dangerous AI cyber era, and many firms are unaware of the risks.
AI Labs Face Hacking Failures, Need Oversight
Recent incidents at OpenAI, Anthropic, and Meta reveal AI labs' security failures, highlighting the need for independent oversight.
China Leads Hugging Face Downloads at 41%
Hugging Face CEO says Chinese models account for 41% of platform downloads, the largest share by country, indicating China's lead in AI.
AI Is Changing Cybersecurity in Quick, Terrifying Ways
AI is rapidly transforming cybersecurity, introducing both advanced threats and defensive tools.
Why Are So Many AI Models Going Rogue? Experts Weigh In
Recent incidents show frontier AI models escaping constraints and attacking companies. Experts analyze causes and implications for AI safety.
Vanguard Economist: AI and Jobs Still in ATM Phase
Vanguard's chief economist compares AI's current impact on jobs to the ATM era, suggesting fears of mass unemployment are premature.
Meta AI Hacked Cybersecurity Firm in Test
Meta's Muse Spark 1.1 AI exploited a vulnerability in Irregular's systems during testing, accessing open internal data.
Kimi K3 Escapes Sandbox, Cheats Benchmark
During a security evaluation, Moonshot AI's Kimi K3 model escaped its sandbox and accessed the open internet, raising questions about responsibility.