How to Tell Whether Your Employer Uses AI to Evaluate Workers

Updated 10 November 2026 · Standard formula, shown below

You can look for signs that software helps score, rank, monitor, or make decisions about employees, then verify those signs through policies and careful questions. No single clue proves that a system uses artificial intelligence, so gather documentation and distinguish automated analysis from ordinary digital recordkeeping.

What You’ll Need

Access to your employee handbook, privacy notice, monitoring policy, performance-review materials, and any relevant workplace technology policy.
Examples of notices, dashboards, review forms, or messages that describe how your work is measured.
A private place to keep dated notes and copies of documents you are allowed to retain. Do not copy confidential business or coworker information.
If useful, a trusted union representative, employee representative, or HR contact to help you ask questions.

Step-by-Step

Define what you are trying to find out. “AI evaluation” can mean different things: software that recommends a performance rating, automated scoring of calls or written work, ranking employees for assignments, or a tool that flags people for review. It can also mean automated monitoring, such as counting keystrokes or measuring time on tasks, without AI. Write down the specific decision or activity you are concerned about, such as promotion, scheduling, discipline, productivity ratings, or a formal review.
Check workplace policies and notices. Search your handbook, privacy notices, onboarding documents, performance-review instructions, and monitoring disclosures for terms such as “automated decision,” “algorithmic,” “machine learning,” “AI,” “productivity analytics,” “automated scoring,” “workforce analytics,” or “decision-support tool.” Check whether the document identifies the purpose, data collected, affected decisions, vendor, retention period, and whether a manager reviews the result. A policy may use broad language, so absence of the word “AI” is not proof that no automated tool is involved.
Look for signs in the evaluation process. Note whether ratings arrive unusually quickly, use identical explanations, include metrics you were not told were collected, or change after a new system was introduced. Look for scorecards, dashboards, alerts, ranked lists, automated summaries, or recommendations displayed to managers. These are reasons to ask questions, not proof of AI: templates and conventional reporting software can produce similar signs.
Separate data collection from decision-making. A system can record hours, calls, tickets, sales, or task completion without evaluating employees. Ask whether software merely stores or summarizes information, generates a score or recommendation, ranks people, or makes a decision automatically. Also ask what role a manager plays: whether they can see the underlying evidence, challenge a score, change the result, and explain the final decision.
Keep a factual record. Record dates, the name or description of the tool if known, what you observed, and the exact wording of relevant notices or explanations. Save material you are authorized to access, such as your own review or a policy copy. Avoid secretly recording conversations unless you have checked applicable workplace rules and local law; recording rules vary. Do not access restricted systems, use another person’s login, or collect coworkers’ personal data to investigate.
Ask a neutral, specific question in writing. You could ask HR, your manager, or the privacy contact: “Does any automated system score, rank, monitor, or recommend decisions about my performance, scheduling, promotion, or discipline? What information does it use, and does a person review and have authority to change the result?” Request the relevant policy or a description of the process. Asking about a concrete decision is more likely to get a useful answer than asking only, “Do you use AI?”
Follow up on how the system works. If the answer is yes, ask what data is included, how often it is updated, which decisions it informs, how errors are corrected, and how you can provide context or challenge inaccurate information. Ask whether the tool evaluates your work directly or flags cases for a human reviewer. A human in the process does not necessarily mean the tool has little influence, so clarify where the score enters the decision.
Compare the explanation with what you can verify. Check whether the described system matches the policy and your review materials. If a rating seems wrong, identify the specific metric, date, task, or context you dispute and provide supporting records through normal workplace channels. Ask for a review by someone who can assess the underlying evidence, rather than relying only on a general statement that the score was generated automatically.
Escalate carefully if you receive no clear answer. Ask the designated privacy, compliance, employee-relations, or HR contact, or consult a union or employee representative if you have one. Keep communications factual and save replies. If the issue involves a consequential decision or your rights, rules differ by location and circumstance; consider a qualified local adviser or the relevant public agency rather than assuming one policy applies everywhere.

Common Mistakes

Treating monitoring as proof of AI. Time tracking, productivity reports, and dashboards may use simple rules or ordinary software. Ask what the system actually does.
Assuming a human reviewer makes the process fully manual. A manager may approve a recommendation while still relying heavily on an automated score. Ask how the recommendation affects the decision.
Relying on rumors or one unusual rating. A strange result can have many explanations. Look for written evidence and request a specific explanation before drawing conclusions.
Using technical terms as if they were interchangeable. Automation, algorithms, analytics, and AI are related but not identical. Ask about functions, inputs, and decision authority, not just labels.
Taking or sharing information improperly. Do not bypass access controls, expose confidential records, or collect coworkers’ information. Keep your inquiry within authorized channels.

Tips

Ask about one process at a time—for example, performance ratings—so the response can be precise.
Use neutral wording and request written clarification. This creates a clearer record without assuming misconduct.
Ask how to correct inaccurate source data, such as a misattributed task or incomplete work log, as well as how to appeal the final rating.
Check when a policy was updated and whether it covers contractors, remote work, or only certain teams; policies can differ across roles.
Keep your own review materials and relevant communications in a secure, appropriate place, following workplace confidentiality rules.

Troubleshooting

Your employer says “we don’t use AI.” Ask whether that means no AI is used at all or only that no decision is made without a person. Describe the function you are asking about—such as automated call scoring or employee ranking—and request the policy that covers it.

You receive a vague answer about “analytics.” Ask what inputs are analyzed, whether the output is a score, flag, or recommendation, who sees it, and whether it affects a specific employment decision. If the contact cannot answer, ask who owns the system or privacy notice.

You cannot find a policy. Request the current monitoring, privacy, or performance-evaluation policy from HR or the appropriate employee portal. Keep the request simple and ask where notices about workplace tools are normally published.

You believe the score contains an error. Identify the disputed item and supporting evidence, then use the stated correction or review process. Ask for the underlying information relevant to your own evaluation if it can be shared. Do not try to inspect systems or records you are not authorized to access.

You are worried about retaliation or a high-stakes outcome. Keep a factual timeline and communicate through established channels. Consider seeking confidential guidance from a union representative, employee adviser, or qualified local professional. Avoid making public accusations before you have verified the facts.

Frequently Asked Questions

Can I tell from a performance score whether AI was used?
No. A score may come from a spreadsheet, fixed rules, a manager, or an AI-enabled tool. Ask what generated it and what information was used.

Does employee monitoring automatically mean AI evaluation?
No. Monitoring can collect or display information without analyzing it using AI or affecting a decision. Ask whether the data produces scores, recommendations, rankings, or alerts.

Can I ask my employer directly?
Yes. Ask the appropriate manager, HR contact, or privacy contact a specific, neutral question about the tool, data, decisions, and human review. The process and any disclosure obligations can vary by location.

What if my employer says a person makes the final decision?
Ask whether the person can review the underlying data, disregard or change the recommendation, and explain the result. Human involvement does not show how much weight the software carries.

Should I secretly record a meeting to prove what was said?
Do not assume recording is permitted. Laws and workplace rules vary. Safer options include taking dated notes and sending a polite follow-up email summarizing your understanding.

What should I do if I think the system got my information wrong?
Point to the specific item, explain why it is inaccurate, provide appropriate supporting evidence, and request correction or review through the employer’s normal process.

Method: the standard formula for this calculation, rounded as each result states. See how the calculators work or report an error.