Insights on AI security and data leaks

Notes from our founder, Firdous Mahmood, on how secrets leak into AI tools, why alignment on the server is not the same as trust on the user's side, and what an AI firewall should be.

Read the latest See EraseAI Firewall

Vee the robot reading by lamplight, with three article pages floating up

Articles

LinkedIn · 1 October 2026

Your team is already pasting secrets into AI chats. Here's the seatbelt we built.

Smart people paste secrets into ChatGPT every day. Not carelessly, just fast: a config file with a cloud key, a customer export. Why that happens, and how EraseAI Firewall catches it at the Send button.

LinkedIn · 24 April 2026

The Missing Layer in AI: A Firewall the User Owns

Everyone is teaching their AI agents to behave. Nobody is giving the user a comfort zone. Why server-side alignment isn't the same thing as user-side trust, and why the firewall should be independent of the model it protects.

Medium

The Three Layers Between a Citizen and an AI Model

A long-form look at the layers of protection, governance and control that should sit between a person and the AI systems they use.

More from Vantward

EraseAI FirewallStops secrets and personal data reaching AI chats. RubikationLive 1v1 cube arena with world and country ranks. CookHubZero-commission home cooking from your neighbourhood. TETHER: PrimalCo-op dinosaur survival. Coming soon to PC.