AI Authority Level Classifier

Free, no signup. Answer five questions about one workflow. We classify by the highest authority level the AI can reach before a human can meaningfully intervene.

1. What does the AI primarily produce in this workflow? A first draft or starting point a human rewrites A summary or rollup of existing material A recommended option or course of action A ranked or filtered shortlist The selected answer or decision

2. Before output is used, what review happens? Full human edit on every use Human checks against source material Human explicitly chooses among options Human sees only what the AI surfaced Human accepts or overrides — often quickly

3. Can the AI trigger action in other systems? No — output stays in a document or chat Rarely — human copies output elsewhere Sometimes — queues or tickets are created Yes — prepares actions pending approval Yes — acts without per-action review

4. What happens to the output over time? Discarded or heavily rewritten if used Becomes the convenient record people cite Becomes what people see — source rarely checked Treated as the decision unless someone objects

5. Who is accountable if the output is wrong? The person who edited and sent it The person who acted on the recommendation Unclear — "the tool decided" or vendor blamed Whoever approved the queued action No clear owner — autonomous path