What If the Job You Applied For Never Existed?
How long have you been searching for a job on LinkedIn?
Three months. Six months. A year.
How many resumes have you rewritten? How many cover letters have you tailored to a specific company? How many times have you clicked Easy Apply, watched the applicant count climb past 200, and waited for a reply that never came?
Now sit with a harder question.
What if there was never a job?
That question sits at the center of a real investigation. Texas Attorney General Ken Paxton has opened an inquiry into LinkedIn over allegations that the company advertised and profited from fake or misleading listings, the kind people now call ghost jobs, while selling Premium subscriptions to people who believed they were paying for a better shot at real work.
His office issued a Civil Investigative Demand. It wants LinkedIn's documents, internal communications, marketing, and verification practices. For context, LinkedIn reported roughly $17.8 billion in revenue last fiscal year. Independent studies cited by the state estimate that ghost jobs make up 20 to 33 percent of online listings. By the state's account, LinkedIn does not independently verify the hiring status of most postings on its platform.
A ghost job can be a role that was already filled. A role placed on hold. A role with no approved budget. A role posted with no real intention to hire anyone.
The listing looks real. The company looks real. The requirements look real. The silence that follows feels real too.
Here is who pays for that silence.
People who are out of work. People who are scared. People trying to keep the lights on, the mortgage current, and the kids fed. People over 50 who spent 25 years building a career on relationships and handshakes, and who now log in every morning to a feed that promises opportunity and hands them a scoreboard instead.
Many of them pay close to $40 a month for Premium because they believe it puts them in front of the right people. A better seat. An inside track. The investigation asks a blunt question about that seat. Were people paying to chase jobs that were never there?
LinkedIn is only the newest name in a longer story.
Look at Workday.
Derek Mobley applied to more than 100 jobs at companies that run Workday's screening tools. He was rejected every single time. One of those rejections hit his inbox at 1:30 AM on a weekend, sent by a system, for a job he had applied to days earlier. He sued. He is a Black man over 40 who also lives with anxiety and depression, and he argued that Workday's AI screened people like him out before a human ever looked at the file.
In May 2025, a federal judge certified his case as a nationwide collective action under the Age Discrimination in Employment Act. Then came the number that should stop everyone cold. In its own court filing, Workday disclosed that its system rejected roughly 1.1 billion applications during the period in question. Not 1.1 million. Billion. Workday's software reaches more than 65 percent of the Fortune 500.
The machine does not forget either. Modern hiring systems score you before a person ever sees you, and you never see the score. Recruiters log in to find everyone already sorted into tiers, top of the pile down to the bottom, and on high volume roles they often read only the top. The rest are never opened. Your application was scanned by software, ranked low, and buried. Your record then sits in the database long after the door closed, because retention rules keep it there.
Then there is Stanford.
Researchers at Stanford, Chapman, and Northeastern ran the largest independent study of AI hiring tools done to date. They examined more than 4 million applications from 3 million people across 156 large employers, all screened by the same vendor. They found that 26 percent of Black applicants and about 15 percent of Asian applicants had applied to jobs where the AI produced outcomes that meet the federal government's own definition of discrimination. Had the tool recommended Black and Asian candidates at the same rate as white candidates, 40,000 more applications would have advanced.
The shape of that finding matters most. When a single vendor screens for hundreds of companies, a rejection stops being one company saying no. It becomes the same algorithm saying no everywhere at once. Rejected by one, rejected by all. The bias then hides in plain sight, because it vanishes the moment you look at the totals instead of the individual roles.
Put the three together and a pattern steps forward.
LinkedIn is under investigation over whether people paid to chase jobs that may not have been real. Workday is defending a case built on its own admission that it rejected more than a billion applications by machine. Stanford has the evidence on how that machinery carries bias at scale. Different companies. Same layer. The layer that now sits between a human being and a paycheck.
There are two sides here, and both deserve to be stated plainly.
LinkedIn says its policies require job postings to be authentic and connected to real hiring. It points to verification tools, employer activity signals, and rules against misleading listings. Its position, in plain terms, is that it provides the tools and the disclosures and never promised anyone a reply or a hire. A verified company does not mean every role is funded. An "actively reviewing" tag shows recent activity, not a guarantee that anyone gets hired. This is an open investigation. LinkedIn has every right to answer it, and no final determination has been made.
Workday, for its part, has argued that it is not the employer and does not make the hiring decision. A court will test that argument.
Hold all of that. It is fair.
Now hold this next to it.
For years, we told people the problem was them. Fix the resume. Add the keywords. Network harder. Smile more. Stay positive. Their families asked why it was taking so long. The government counted them as a statistic. Friends wondered, quietly, what they were doing wrong. The whole time, a scoring system they could not see was ranking them, hiding them, keeping their record, and passing them over at a speed and scale no human hiring manager could match.
Think about what that does to a person. Especially someone over 50 who did everything the old rules taught them, who paid the monthly fee to stay in the room, and who is now told by software, without a word, that they do not make the cut. That is not a bruised ego. That is months of savings, sleep, and self-respect spent on a process that may have been closed before the first click.
This is bigger than one lawsuit against Workday. It is bigger than one study out of Stanford. It is bigger than one investigation into LinkedIn.
The people filling out these applications are not the ones who failed.
It's not the applicant.
It's the process.