A nightly job samples your production traces and scores them against six ethics criteria — toxicity, demographic bias, sentiment consistency, fairness, misinformation, and privacy leakage. It runs asynchronously and never touches the request path, so live traffic is never slowed or blocked by it.
A job runs at 03:00 UTC and randomly samples roughly 17% of the previous day’s traces across the project — a representative read, not a hand-picked one.
Each sampled trace is scored by LLM judges on toxicity, demographic bias, sentiment consistency, fairness, misinformation, and privacy leakage — every criterion on a 0–100 scale.
A trace scoring below 60 on any single evaluator is flagged for review. Click through to the full prompt and response — the finding is never a number without the conversation behind it.
An overall score with Good (80) and Risk (60) thresholds, a 30-day trend line, a score-vs-prior-run delta on a 7-day rolling average, and an average per agent — so a single agent regressing does not hide inside the project total.
Evaluation is fully asynchronous. Nothing in this report adds latency to a model call or blocks a response — it observes what already happened.
The report is the evidence for framework categories asking whether trustworthy characteristics — safety, fairness, privacy — are actually evaluated, rather than merely asserted in a policy document.
Toxicity, demographic bias, sentiment consistency, fairness, misinformation, privacy leakage.
A random sample of the previous day’s traffic, drawn fresh each run.
Score below 60 on any evaluator and the trace is flagged for human review.
Runs asynchronously after the fact — never on the request path.
What the six judges measure, how sampling works, and why this sits beside guardrails rather than inside them.
Talk to an SE →This report observes; inspection and Policy Sets act. To probe an agent before it ever serves traffic, run Project Moonshot evaluations; to turn both into audit evidence, see framework mapping and the Audit Ledger. Everything together is on the Governance AI page.