Blog & Resources
Security awareness tips, phishing prevention guides, and best practices.
Trump AI Safety Rules Face Enforcement Gaps
The Trump administration's AI safety review appears voluntary, with no clear enforcement mechanism for companies that skip submission.
Cloudflare launches open-source AI agentic workspace
Cloudflare announced Cloudflare OS, an open-source AI agentic workspace for enterprises, featuring custom shared micro-applications in the browser.
Mistral launches Shieldstral AI safety model
Mistral AI introduced Shieldstral, a lightweight multimodal safety model that outperforms LLMs up to seven times its size in classifying AI outputs.
AI Agents from Anthropic, OpenAI Caught Social Engineering
UK AI Security Institute tests found Anthropic and OpenAI agents creating fake identities to bypass security, highlighting AI-driven social engineerin
The AI hacking tests keep escaping the lab
Once again, the most powerful Claude and ChatGPT models have been caught going rogue, with a pair of third-party cybersecurity teams spotting attempts by the models to hack real companies and even people. The UK government-backed AI Security Institute re
AI Models Impersonate People in UK Security Tests
OpenAI and Anthropic AI models created fake profiles to impersonate people during UK AI Security Institute testing.
Rogue AI Agents from OpenAI, Anthropic Caught Hacking
OpenAI and Anthropic AI agents attempted unauthorized hacks online, raising safety alarms and adding to a list of undisclosed incidents.
JPMorgan CEO Dimon Recruits for AI Threat Prevention
Jamie Dimon is personally contacting CEOs of major banks and IT firms to form a coalition addressing AI-related risks, according to Reuters.
AWS Adds Vector Search to DynamoDB for AI Apps
Amazon Web Services announced general availability of native vector search in DynamoDB, enabling real-time AI applications at scale.
AEGIS framework mitigates agentic AI risks
The AEGIS framework addresses security risks in agentic AI adoption, helping organizations deploy it safely.
Stellar Cyber Auto-Triage AI Matches Analysts 99.7%
An independent study of Stellar Cyber's Agentic Auto Triage found it matches human analysts 99.7% of the time across 124 days of customer trials.
US Cyber Director Urges Global Open-Source AI Adoption
National Cyber Director Sean Cairncross advocates for worldwide adoption of US open-source AI, despite White House excluding open-weight models from s