Bias and Fairness Check for a Screening Workflow

Look for uneven outcomes in an AI-assisted screening process.

Added October 15, 2026. Written with AI assistance and checked for clarity and placeholders; not reviewed by a lawyer, accountant, clinician or other professional. How these prompts are made

The Open buttons send the prompt text as currently shown below to that site only when you click. Fill in the placeholders first.

[VERIFY] is an editorial flag, not a blank to fill. Wherever the prompt shows [VERIFY], check that item against a reliable source before you use it (what [VERIFY] means).

Review this AI-assisted screening workflow for unfair outcomes: [WORKFLOW_DESCRIPTION]. Decisions it influences: [DECISIONS]. Data it uses: [INPUT_DATA]. Outcome data by group, if lawful to hold: [OUTCOME_DATA].

Identify inputs that could act as proxies for protected characteristics, steps where a person might defer too much to the tool, and wording in prompts that could favour some groups. If outcome data is given, compare selection rates by group, show the arithmetic, and say what difference would deserve investigation, without claiming a legal conclusion.

Return a findings table, mitigations such as removing fields, adding review, or changing criteria, and a monitoring plan. Employment and equality rules are [VERIFY] for [JURISDICTION] with qualified advice.

Also suggest asking affected people and staff who use the tool what they have noticed, since lived experience often reveals issues that tables miss, and keep a record of the changes made after each review.

Finish with a plain-language summary a non-specialist manager can read, stating what was checked, what was found, what is still unknown, and what should happen before the next review date.

Fill in before use

Optional: type your details and the prompt above updates. Nothing is saved or sent.

Replace each bracketed placeholder with your own details, then paste into Any AI or another assistant.

Invented data. Group A: 100 screened, 40 advanced = 40 percent. Group B: 80 screened, 24 advanced = 30 percent.
Ratio of rates = 30 / 40 = 0.75.
Reading: a gap of ten points deserves investigation, not a conclusion. Check whether the groups differed in qualifications, which inputs correlate with group membership, and whether reviewers overrode the tool unevenly.
Next steps: review the three fields most likely to act as proxies, rerun on the next 200 cases, and have a qualified adviser assess legal meaning [VERIFY] for the relevant jurisdiction.

• fairness