The Age Discrimination Machine

She applied for 143 jobs over eleven months. She has an MBA, twenty-eight years of experience in brand marketing, and a track record of growing revenue at three different companies. She knows how to do the work. She has done it well for nearly three decades.

She got seven interviews. All of them stalled after the video call.

One interviewer asked when she graduated from college. Another asked if she would be “comfortable reporting to someone younger.” A third said the team was looking for someone with “a fresh perspective,” which meant nothing and everything. The rejections, when they came, cited culture fit. Or said nothing at all.

She is fifty-seven years old. She is too experienced. She is too expensive. She is too old. No one says this explicitly. No one has to. She has learned to read the silences.

How Discrimination Actually Works

#

Age discrimination does not arrive with a note explaining itself. It operates in the spaces between what is said and what is done, in decisions made before a human ever reads a resume, in language designed to signal preference without leaving evidence.

The hiring funnel has multiple stages, and older workers drop out disproportionately at the earliest ones. Resume screening, increasingly automated, filters applications before any person evaluates them. Initial phone screens with HR gatekeepers assess “fit” based on voice, energy, and answers to questions designed to reveal age without asking it directly. By the time a candidate reaches a hiring manager, the pool has already been shaped by decisions no one can see.

“Culture fit” is the language that does the work. Job postings seek candidates with “high energy,” “digital native” skills, “startup mentality,” or “fresh perspectives.” These phrases do not mention age. They do not need to. Everyone understands what they mean. A fifty-seven-year-old reading such a posting knows she is not the target audience, even if she could do the job better than anyone else who applies.

Compensation assumptions filter out older candidates before salary is ever discussed. Employers assume experienced workers will demand higher pay. Rather than negotiate, they screen them out. The assumption is often correct, since experience commands higher market rates, but the result is that older workers never get the chance to decide whether they would accept less.

“Overqualified” is the polite version. It means you have too much experience for this role, which means you are too old for this role, which means we have found a legal way to reject you for an illegal reason. No one says this. Everyone knows it.

The deepest problem is invisibility. When you apply for a job and hear nothing, you have no evidence of what happened. You cannot prove discrimination if you cannot see where in the process you were eliminated. The system is designed to leave no fingerprints.

The Algorithm Problem

#

Resume screening was once a human activity. It is now largely automated. Applicant tracking systems score resumes against job requirements. Machine learning models predict which candidates will succeed based on patterns in historical hiring data.

The problem is that historical hiring data reflects historical discrimination. If a company has consistently hired younger workers for a given role, the algorithm learns that youth predicts success. It then screens out proxies for age: graduation years from too long ago, decades of experience that suggest seniority, technology skills that were current in 2005 but have been superseded, employment gaps that might indicate caregiving responsibilities.

The algorithm does not know it is discriminating. It is optimizing for patterns. The patterns encode bias.

This is not theoretical. Meta settled claims with the Department of Justice after its advertising tools allowed employers to target job postings by age, excluding older workers from ever seeing the positions. That case was overt. Most algorithmic discrimination is invisible, embedded in systems whose internal logic is proprietary and whose outcomes are never audited.

The Equal Employment Opportunity Commission issued guidance in 2023 clarifying that employers are liable for discriminatory outcomes from AI hiring tools, even if the discrimination was unintentional. But enforcement requires detecting the discrimination first. Applicants who are screened out before a human sees their resume have no way of knowing why. They cannot file a complaint about bias they cannot observe.

Audit studies, where researchers submit identical resumes with different ages implied, show the disparity. In some fields, callback rates for older applicants run 30 to 40 percent lower than for younger candidates with equivalent qualifications. The discrimination is measurable in aggregate. It remains invisible in individual cases.

What the Law Provides

#

The Age Discrimination in Employment Act has protected workers forty and older since 1967. It prohibits discrimination in hiring, firing, compensation, and terms of employment. It applies to employers with twenty or more employees. It is enforced by the EEOC.

On paper, the protection is substantial. In practice, it is hollow.

The Supreme Court’s 2009 decision in Gross v. FBL Financial Services raised the burden of proof for age discrimination claims. Plaintiffs must now show that age was the “but-for” cause of the adverse action, meaning the outcome would have been different if the plaintiff were younger. This is a higher standard than applies to claims under Title VII, which covers race and sex discrimination. Under Title VII, a plaintiff can prevail by showing that a protected characteristic was one factor in the decision, even if other factors contributed. Age claims do not get this flexibility.

The practical effect is that proving age discrimination requires either a smoking gun (an email saying “we don’t want anyone over fifty”) or a pattern so stark that no alternative explanation is plausible. Neither is common. Employers who discriminate know how to do it quietly.

The EEOC receives roughly 12,000 to 15,000 age discrimination charges annually, about 20 percent of all charges filed. Most result in no-cause findings or small settlements. Litigation is expensive, slow, and uncertain. Attorneys take cases on contingency only when the evidence is overwhelming and the damages are substantial. For the marketing director who was screened out of 143 jobs, there is no case to bring. She has frustration, not evidence.

Some states offer stronger protections. California, New York, and New Jersey have lower thresholds, broader coverage, or additional remedies. But most states follow federal minimums, and even strong state laws cannot overcome the fundamental problem: discrimination that leaves no trace cannot be proven.

What Could Change

#

Legislative reform exists on paper. The Protecting Older Workers Against Discrimination Act would restore the mixed-motive standard for age claims, aligning the burden of proof with Title VII. It has been introduced repeatedly. It has never passed.

Algorithmic accountability is emerging in fragments. Illinois requires employers to notify candidates when AI is used in video interview analysis. New York City mandates bias audits for automated employment decision tools. These are narrow measures, limited in scope and inconsistently enforced. Federal action is sparse.

Some employers are reconsidering their assumptions, not from altruism but from necessity. Labor shortages in certain sectors make it harder to ignore older talent pools. Research on age-diverse teams suggests they outperform age-homogeneous ones on complex problems. Demographic pressure may accomplish what legal pressure has not.

For individuals, the available actions are defensive. Document everything. Ask for feedback in writing. Consult employment attorneys early. Network aggressively, since referrals bypass some screening. Remove graduation dates from resumes, though this signals its own message. None of this fixes the structural problem. All of it may help at the margin.

What Remains

#

The law says you cannot be rejected for a job because you are fifty-seven. The reality is that you can be rejected for a job because you are fifty-seven, and unless someone writes “too old” in an email, you will never prove it.

The hiring process is a black box. The algorithms are trained on data that reflects the biases of the humans who made the previous decisions. The language of “culture fit” and “fresh perspective” does the work that explicit exclusion cannot. The burden of proof ensures that most discrimination goes unchallenged.

What would change this is not better enforcement of existing law, though that would help. It is a shift in how organizations think about experience: seeing it as an asset rather than a cost, understanding that the judgment developed over decades cannot be replicated by enthusiasm alone, recognizing that the speed of youth and the pattern recognition of experience are complements rather than substitutes.

That shift is not happening fast enough. The fifty-seven-year-old marketing director is still applying. She has stopped counting the rejections.