Human oversight
Last updated: 14 July 2026
Regulators across the EU, UK, and US converge on one requirement: a human must be able to understand, question, and overrule the machine. A rubber-stamp click does not meet that bar. This page describes how Stori is designed so that human review is meaningful — capable of catching the machine being wrong — rather than ceremonial.
What makes review meaningful here
Receipts, not verdicts
Every interpretation, score, and ordering ties back to underlying evidence — the transcript passage, the quote, the context. A reviewer never has to take the AI's reading on faith; they can check it against what the candidate actually said and reject it.
Override, built in and captured
Disagreeing with the machine is a supported action, not a workaround. Overrides are captured in the product, which gives an employer a record that its review is genuine — evidence several regimes expect a deployer to be able to produce.
One input among several
Stori's output is presented as decision support alongside everything else an employer knows — never as an auto-decision, and it is not designed to be used as the sole basis for rejecting anyone.
Stated confidence and abstention
Each facet score carries a confidence, and thin evidence yields “insufficient signal” rather than a fabricated number. A reviewer always knows how much weight an output can honestly bear.
Plain-language logic
Stori states in plain language what its outputs are and are not — what it reads, how it orders evidence by strength and checkability, and what it does not claim — so the person reviewing can interpret the output appropriately.
The candidate's side of oversight
- Candidates own their profile and their evidence — the interview is theirs, built once and portable, not an employer's verdict that follows them.
- Where an employer's process gives candidates a right to explanation, human intervention, or contest, Stori's source-linked outputs are designed to make that answerable — the explanation of “the role the AI played and the main elements” can be given from the receipts, not from a bare score.
- Candidates who need an adjustment to complete a Stori interview — extra time, restarts, or an alternative format — can request one at support@onestori.com, stated in the interview screen before they start.
What we do not claim about oversight
We do not claim that human review is a liability shield, and we do not market Stori as reducing an employer's legal exposure. Courts are actively testing the “human made the final call” defense in AI-hiring litigation, and several statutes attach duties to output that substantially assists a decision whether or not it is final. What we claim is narrower and checkable: every output shows its work, so the human in the loop can genuinely exercise judgment — which is what the oversight requirements at the center of these laws actually demand.
This document describes how Stori is designed. It is not legal advice and does not determine whether any specific employer deployment is lawful. Questions from legal or procurement teams: support@onestori.com.