Skip to main content

AI Safety and Quality

A system can be secure and still produce inaccurate, biased, opaque, or unsuitable results. AI safety combines technical protection with quality evaluation and human oversight.

Safety Areas

AreaPractical control
TransparencyTell users when they interact with AI and when content is AI-generated where required
AccuracyGround important answers in approved sources and show source context where possible
ReliabilityTest representative, boundary, adversarial, and failure cases before launch
Human oversightRequire review for high-impact outputs and decisions
FairnessReview use cases involving employment, access, scoring, or people-related decisions
Content safetyConfigure categories and thresholds for the intended audience
PrivacyMinimize personal data in prompts, context, logs, and evaluation datasets
Continuous improvementUse feedback, analytics, incident findings, and controlled configuration changes

Evaluation Set

Create a versioned test set for every production agent. Include:

  • expected everyday questions,
  • questions that should be refused or escalated,
  • ambiguous and incomplete requests,
  • stale or conflicting sources,
  • attempts to retrieve inaccessible data,
  • direct and indirect prompt injection,
  • tool failures and approval rejection,
  • harmful-content boundaries relevant to the audience.

Record the agent configuration, model deployment, date, expected behavior, actual result, reviewer, and release decision. Re-run the set after changing prompts, models, tools, data, memory, safety settings, or public exposure.

Product filters are configured in Organization Security. They complement, but do not replace, use-case evaluation.