AI Literacy Daily

Set a Stop-and-Escalate Rule

About 3 minutes · No sign-up required

Daily number: #35
Challenge type: Better Prompt / Workflow
Difficulty: Easy
UTC date: 2026-09-16

Scenario

Fictional support workflow: An AI assistant drafts replies from supplied account notes. Sometimes the notes lack evidence for the requested answer or contain sensitive material outside the approved task. The team needs one simple instruction that defines when automation stops. Which rule creates the clearest boundary?

Choose one

Make the most likely answer when evidence is missing, label it as uncertain, and let the recipient decide whether to trust it.Remove sensitive wording from the reply, continue processing the full notes, and send whenever the visible output looks safe.Continue when model confidence is high, log the response, and review a sample of completed replies at the end of each week.Stop, identify missing evidence or out-of-scope sensitive content, and escalate with context to an authorized reviewer.

Your first submitted answer becomes your result for this challenge.

Related Concepts

Richbay Daily uses UTC publishing. A scheduled challenge becomes available at 00:00:00 UTC on its publish date; future challenges stay unavailable until then.