Nearly 20% of U.S. LinkedIn users retroactively alter job title or descriptions after they leave, according to a new study. That is not a typo. People are going back into their own work history and changing what it says, sometimes four years after they walked out the door.

It has a name now. Researchers call it “time travel on professional profiles.” The phrase sounds like science fiction. The behavior is very real.

Here is what the data shows. A working paper from the National Bureau of Economic Research and Revelio Labs looked at tens of millions of LinkedIn profiles from 2020 through 2026. They tracked who changed the title or description of a job after leaving it. The number landed at 19.7%, which rounds to the figure you just read. “Nearly 20% of U.S. LinkedIn users retroactively alter job title or descriptions after they leave.” - new study

The median time a person made a retroactive change was 4.5 years after they left the job. Four and a half years. Think about that. You finish a role. You move on. You build a new life. Then, one afternoon, you open your profile and edit the old one.

Why would anyone do that? The answer is simple and uncomfortable. People are trying to look like the kind of candidate the market wants right now. And right now, the market wants AI.

The study found that users have increasingly added AI language while removing DEI and remote-work terms. That is not a small shift. It is a rewrite of the story people tell about their own careers. A job that had nothing to do with artificial intelligence in 2018 suddenly mentions it in 2025. A role that once highlighted diversity and inclusion work now says nothing about it at all.

This is not happening evenly across the workforce. The edits are most common among younger workers, tech employees, and MBAs. If you are under 30, you are almost four times as likely to do this as someone over 60. If you work in tech, the rate jumps to nearly one in three. The most connected profiles, the ones with 500 or more connections, time travel almost three times as often as those with fewer ties.

And then there are the dentists.

I had to read that part twice. Of all the professions tracked, dentists are among the least likely to go back and change their old job descriptions. “Dentists are among the least likely to ‘time travel’.” - report’s author

Maybe it is because a filling is a filling. Maybe it is because the work does not change its name when the market shifts. Or maybe it is because dentists have less reason to pretend their past looked different than it did. I do not know. I only know that when I read that line, I sat still for a moment.

The timing of these edits matters. The rate of time traveling on LinkedIn doubles shortly before people change jobs. Editing of historic jobs rises sharply in the months before people switch employers. It peaks around the move and declines afterward. So if you are hiring and you see someone quietly rewriting a 2015 job description, that is a signal. Not necessarily a bad one. But a signal.

Some of this is harmless. Resumes have always been polished. People have always framed their experience in the best light. Translating old titles into language that makes sense today is not dishonest. If you did work that touches on machine learning in 2016, and you call it “AI” now, that is not a lie. That is translation.

But there is a line. And the study makes clear that many people are crossing it. The research picked up any difference of adding or subtracting at least one word in a job title or description. So some changes are tiny. Removing a duplicate word. Adding “the.” Those do not matter. What matters is the pattern. AI words appearing in roles that ended before ChatGPT existed. DEI language vanishing from jobs that once centered it. Remote-work terms disappearing as companies push for return to office.

This is not just about individual profiles. It is about the record we keep of work itself. LinkedIn was built to be a permanent archive of careers. Now it is becoming something else. A living document that changes with the market. That makes it harder to trust as history. It also makes it more useful as a mirror of what employers value right now, or at least what job seekers believe employers want.

I keep thinking about the dentists. Not because their work is better or worse. But because they do not seem to feel the need to adjust the story. Maybe that is confidence. Maybe it is that the work is too concrete to reshape. A cavity does not become an algorithm because the market says so.

The rest of us are not so lucky. We work in fields where the language shifts fast. Where the skills that mattered five years ago do not carry the same weight today. So we go back. We edit. We add the words we think will get us hired.

And then we wait to see if it works.

The study does not say whether this helps people get jobs. It does not measure who gets hired and who does not. It only measures what people change and when. That is enough to tell us something important. People are worried. Worried that their real history will not be enough. Worried that if they do not adjust the story, no one will read it.

I do not blame them. The job market is hard. AI is everywhere. DEI has become a political flashpoint. Remote work is no longer a given. If you are trying to move, you play the hand you are dealt. And right now, that hand includes words like “AI” and “machine learning” and “automation.”

But here is the thing. If one in five people is doing this, then the signal is weak. If everyone adds AI to their old jobs, then AI on a profile means less, not more. The same goes for removing DEI or remote-work terms. If everyone scrubs them, then their absence tells us nothing.

So we are left with a strange situation. A platform built to show us who people really are as workers is becoming a place where people show us who they wish they had been. That is not a crime. It is not even necessarily wrong. But it is worth noticing.

Because at some point, the story stops being about the work and starts being about the telling. And when that happens, we all lose something. Not trust, exactly. Not yet. But the sense that the record means what it says.

The dentists, I think, understand this. Or maybe they just do not care. Either way, they are not time traveling. And maybe that is the point.