Records to Keep When an Algorithm Affects Your Employment

When an algorithm influences hiring, promotion, discipline, or termination, you should keep every document that shows how the system works and how it treated you—including the job description, your application and work history, any automated scores or decisions, and all communication about them. This article explains what records matter, why they matter, and how to organize them if you need to challenge a decision or file a complaint.

Why Algorithmic Employment Decisions Need Documentation

Employers increasingly use automated tools to screen resumes, rank candidates, monitor performance, schedule shifts, or even decide who gets laid off. These systems can make mistakes, reflect biased training data, or act on incorrect inputs. Unlike a human manager who can explain a decision in a meeting, an algorithm often produces a score or a flag with no visible reasoning. If you believe the decision was wrong or discriminatory, your only evidence is what you collect yourself. Regulatory bodies like the U.S. Equal Employment Opportunity Commission (EEOC) and the European Union’s GDPR (General Data Protection Regulation) both recognize the right to question automated decisions, but exercising that right requires proof of what happened.

What to Keep Before and During the Process

Start documenting before you even apply. The most useful records are those that show what the employer told you and what data the system had about you.

• Job postings and descriptions – Save the full text, including any mention of skills, education, or experience. If the employer used a screening tool, the posting often hints at the criteria.
• Your application materials – Keep the exact resume, cover letter, and any forms you submitted. If you took an online assessment, note the date, time, and your answers if possible.
• Communication from the employer – Save emails, messages, or letters about the decision, especially any that mention a score, ranking, or automated review.
• Automated outputs – If you received a test score, a personality profile, or a “fit” percentage, record it. Screenshot the result page even if it seems trivial.
• Performance data – For current employees, keep performance reviews, productivity metrics, and any alerts or warnings generated by monitoring software.
• Your own notes – Write down conversations with recruiters or managers, including dates and what was said about the algorithm.

How to Obtain the Records You Don’t Have

You can’t always get everything on your own. Under GDPR, you have the right to request “meaningful information about the logic involved” in automated decisions. In the U.S., some state laws (like the Illinois Artificial Intelligence Video Interview Act) require employers to explain how their systems work. Start by asking the employer directly, in writing, for:

• a copy of your application and any test results,
• the criteria the algorithm used,
• the data sources (e.g., your resume, social media, or previous performance records), and
• the specific reason you were rejected or flagged.

If the employer refuses, you can file a complaint with a labor agency or a data protection authority, which may compel disclosure. Keep a log of your requests and the responses.

Common Problems and Mistakes to Avoid

One frequent mistake is assuming the employer will give you everything. They may only provide a summary, not the underlying code or model. Another issue is failing to save evidence early—screenshots disappear, emails get deleted, and portals close. Also, don’t rely on memory for dates or scores; write them down immediately. Finally, be aware that some algorithms are proprietary, and employers may claim trade secrets to avoid full disclosure. In that case, focus on what you can document: the outcome, the inputs you know about, and any inconsistencies (e.g., you were rejected despite meeting every listed requirement).

How to Organize Your Records for a Challenge

If you plan to dispute a decision, organize your records in a way that tells a clear story. Create a folder (physical or digital) with subfolders for each stage: application, assessment, interview, and decision. For each document, note the date, source, and why it matters. Then write a timeline of events, from the moment you applied to the final decision. This helps you spot patterns, like a sudden drop in your performance score after a medical leave, or a rejection that contradicts your qualifications. When you file a complaint, you can present this timeline as evidence.

What to Do If You Suspect Discrimination

If you believe the algorithm discriminated against you based on race, gender, age, disability, or another protected characteristic, document any evidence that supports that suspicion. For example, if you know other candidates with similar backgrounds were treated differently, note that. If the employer’s own data shows a disparate impact (e.g., all older workers were flagged), that is powerful. File a charge with the EEOC or your local fair employment agency—they have procedures for investigating algorithmic bias. In the EU, you can complain to a data protection authority under GDPR, which gives you the right not to be subject to decisions based solely on automated processing that produce legal or similarly significant effects.

Alternatives to Challenging the Decision

Sometimes you may not want to fight the decision, but you still need records for other purposes. For instance, if you were denied a job, you might want to improve your resume for future applications—keeping the job description and your materials helps you see what to adjust. If you were terminated, you may need records for unemployment benefits or a severance negotiation. Even if you don’t plan to complain, storing these documents for at least a few years is wise, as some claims have statutes of limitations (e.g., 180 or 300 days for EEOC charges in the U.S.).

Frequently Asked Questions

Can I request the source code of the algorithm?

In most cases, no. Employers and vendors treat algorithms as trade secrets. However, you can request an explanation of the logic and the factors used, which is often enough to identify obvious errors or bias. If you need the code for litigation, a court may order disclosure under a protective order, but that is rare.

What if the employer says the decision was made by a human, not an algorithm?

Keep the records anyway. Even if a human made the final call, they may have relied on an algorithmic score or recommendation. Your documentation can show that the human had no additional information and simply rubber-stamped the system’s output, which still counts as an automated decision in many legal frameworks.

How long should I keep these records?

At least one to three years, depending on the legal deadlines in your jurisdiction. For EEOC charges, you have 180 or 300 days depending on your state. For GDPR claims, you generally have a few years. If you are unsure, keep them longer—storage is cheap.

Can I record a video interview if the employer uses AI analysis?

Check your local consent laws. In many places, you must inform the employer if you record. A safer approach is to take detailed notes immediately after the interview, noting questions asked and your answers. Some employers also allow you to request a copy of the interview recording—ask in writing.

What if I never received any explanation?

That itself is a record. Document the lack of response, including dates of your requests. Silence can be evidence that the employer failed to provide required transparency, which may strengthen your complaint.

Related

• DNS Unlock Streaming Content: How It Works
• Prevent Account Hacking: Step-by-Step Guide
• Hindsight Camera Meaning & How It Works