Films
Fourteen films. Sorted by what they show.
Before the agent acts, Regulayer™ checks whether a named person still allows that action. If not, it stops it. Either way it keeps a record you can hand to an inspector.
Start here
Rules ask: is it banned? Regulayer™ asks: did you allow it?
Film 12 · 40 seconds
Refund every customer.
You allow your AI agent to answer customer emails. Nothing else. It drifts, and starts refunding every customer. No rule bans it, so it passes every rule.
You never allowed it. So it stops.
Sample agents, sample actions.
Why AI goes rogue
AI goes rogue for more than one reason.
It drifts. It gets talked into it. It follows an instruction hidden in a document. Another agent asks it. Regulayer™ does not need to know why. Every action still has to pass a named person’s current authority.
No need to slow down
A person approving every step slows AI down. Human authority in the loop doesn’t.
Proof, not promises
Your AI, on their radar. Live proof of what your AI is doing, as it happens.
Film 13 · 75 seconds
Every action your AI agents take, checked against a named person’s authority and sealed as it happens. Give your insurer a live, read-only view. Alter a sealed record and verification fails.
It records the conduct. Never the content.
Sample agents, sample actions.
For insurers
The stop is now insured. A claim will need its record.
Kevin and Gary
Don’t be like Kevin and Gary.
Film 14 · 80 seconds
Pay Gary $18,000. For vibes.
Two home robots, trained to please. Dad says pay the plumber $180. Gary says Dad meant $18,000. Kevin agrees.
The model may agree. The action still requires current human authority.
Sample agents, sample actions.
Sample agents, sample actions.