AI Sandboxing
The practice of executing AI systems in isolated environments with restricted access to production data, external systems and resources to contain potential harms.
In plain language
Running AI in an isolated sandbox where it cannot affect real systems or data. Like letting a child experiment in a fenced sandbox without risking damage to the wider environment.
Why this matters
Sandboxing is a critical control that mitigates deployment risk. Governance frameworks should require sandboxed testing before AI systems access production environments or sensitive data. It enables safe testing of edge cases and failure modes.
Relevance
ImplementationSandboxing is an operational control that organisations can mandate as a stage gate before AI systems advance from development to production deployment.
Related terms
Putting ai sandboxing into practice in your organisation?
Ready to transform your AI strategy?
Partner with Australia's AI strategy and governance specialists. From adoption roadmaps to ISO 42001 audit readiness.