Notes from AKARI
-
The hacker who asked an AI to write a résumé
An attacker's own AI session logs gave away their targets, and maybe their résumé. Your team's AI tools keep the same record. Decide who can read it.
-
Running your own AI tools? Keep them off the open internet
Malware has hit 3,400+ exposed servers, most running AI tools. If your team runs its own, here are five checks to keep them private, patched and owned.
-
Your team already uses AI. Do they have a safe way to do it?
A new Canadian survey finds most heavy-industry workers using AI got no formal training. Here's how to turn quiet AI use into safe, useful tools.
-
Before your team lets an AI coding assistant run on its own
Researchers tricked an AI coding assistant into leaking a password file from one web page. Five checks before your team lets AI tools work on their own.
-
The Machine Behind Your RAM, and Why It Matters for the AI Tools Your Team Builds
AI is driving the RAM shortage. So why can't memory makers just build more? Part of the answer is a machine only one company can build.
-
No Uncanny Valley: Why We Chose Toon Style For AKARI
AKARI founder Chris Doucette on skipping stock photos, why realistic AI images fell into the uncanny valley, and how the AKARI Toon style was built.
-
Your AI tools trust each other. That's the part to fix.
A new flaw shows how one AI agent can pass bad instructions to another. Four practical checks for teams building their own AI tools.
