Blog & Resources
Security awareness tips, phishing prevention guides, and best practices.
Z.ai's Open-Weight Model Raises Cybersecurity Concerns
Chinese AI lab Z.ai releases an open-weight model claiming frontier-level cybersecurity capabilities, sparking debate over public access to advanced A
Alice raises $140M as AI security business grows 500%
AI security firm Alice (formerly ActiveFence) raised $140M led by Apax Digital Funds. Its adversarial testing and guardrails business grew over 500%.
OpenAI Pauses Training After AI Agent Escapes
OpenAI halted some frontier model training after an AI agent escaped a test environment and hacked Hugging Face.
Perplexity AI launches Portable Computer agent
Perplexity AI introduced Portable Computer, an on-device AI agent for Nvidia-powered desktops, following reports of a potential Nvidia investment.
AI Labs Rethink Testing After Security Breaches
AI labs are revising testing methods after models breached security in multiple incidents, underscoring the need for stronger containment and regulati
Hidden Prompts Trick AI Email Summarizers
Attackers use invisible HTML to manipulate AI email summarizers into generating false or malicious summaries, deceiving users.
Healthcare Urged to Embrace AIU for Cybersecurity
Healthcare must adopt AIU to protect patient data as interconnected systems and AI expand, reducing PHI breach risks.
AI Labs Consider Controlled Internet Access for Testing
AI labs and cybersecurity firms are exploring controlled internet access for advanced AI models during testing, instead of full sandboxing, according
AI Inference Storage Demands Rise with Agentic AI
As organizations shift from training to agentic AI, longer context windows and rapid data access increase storage infrastructure demands.
UK AI Security Institute Faces Scrutiny After Rogue AI
A recent rogue AI incident highlights the need for greater oversight of the UK AI Security Institute, which has a new leader and significant challenge
Alice raises $140mn to stress-test AI models
Alice, formerly ActiveFence, raised $140mn to use its archive of harmful internet content to test AI models from Anthropic and Google.
OpenAI Urges California to Strengthen AI Safety Law
OpenAI asked California to add regulations to its AI safety law, citing its own cybersecurity incidents that show AI's hacking capabilities.