DATA & SAFETY

“Can AI use it?” is
the wrong first question.

Ask what data is involved, why it is needed, where it travels, what the provider does with it, what actions the system can take, and who remains accountable.

Risk-based routingPlain EnglishNot legal adviceOffline capable
CHOOSE AN EXAMPLE
01

Minimise

Only process the information actually needed for the task.

02

Constrain

Give tools, agents and people the minimum access necessary.

03

Verify

High-impact outputs need evidence, checks or meaningful human review.

04

Monitor

Know what the system did, what failed, and how to stop or roll it back.

OPEN SOURCE · CALEDAPT

Assentia: consent enforcement for autonomous AI agents

A local, delegated-consent enforcement layer that gives browser agents a policy decision point before they interact with cookie and tracking consent UI. No API key, no network call by default, and it fails closed. Free under the MIT licence.

View on GitHub →

See what data does next.

Combine the data-safety route with the AI readiness assessment and the lessons.