There's a specific, quantifiable trend researchers have started calling "time travel": job seekers going back into old, closed-out LinkedIn job entries, years after leaving those roles, and adding AI-related language that wasn't part of the job at the time. A July 2026 NBER working paper puts real numbers on it for the first time, and the findings are worth understanding whether or not you've ever edited an old entry yourself, because recruiters, applicant systems, and even AI hiring tools are increasingly aware of the pattern.

What the study actually found

Economists Nicholas Bloom, Gideon Moore, Lisa Simon, and Caelan Wilkie-Rogers analyzed 29.4 million U.S. LinkedIn profiles for the NBER working paper "Time Travel on Professional Profiles." They found that 19.7% of established LinkedIn users retroactively edit the title or description of a job after they've already left it, typically around four years after the fact on average. The edits cluster heavily around career transitions, meaning people tend to revise old job history right when they're actively job searching, which is exactly when a hiring manager or recruiter is most likely to be looking closely at that profile.

AI keywords are driving most of it

The single biggest category of retroactive edit is AI-related terminology. Mentions of terms like AI, GPT, LLM, and "artificial intelligence" being added to old, already-departed job descriptions rose more than sixfold since ChatGPT's late-2022 launch. The researchers estimate that a snapshot of LinkedIn profiles taken today would overstate how much genuine AI skill existed back in 2022 by roughly 30%, purely because of these retroactive additions. The behavior isn't evenly spread: workers under 30 edit past roles at more than double the rate of workers over 60 (38.1% versus under 10%), and technology and information-sector workers show the highest edit rates of any industry, alongside product, UX research, and AI-leadership job titles specifically.

Worth being fair about: not every edit is dishonest. The researchers are careful to note that adding AI language to an old role doesn't automatically mean someone is fabricating experience — some people are genuinely just re-describing work they did using terminology that didn't widely exist yet when they wrote the original entry. The problem is that from the outside, a real update and a padded one look identical, and the sheer volume of edits happening right around job transitions is exactly the pattern that makes recruiters and algorithms treat all of it with more suspicion, not less.

Bloom, Moore, Simon & Wilkie-Rogers, "Time Travel on Professional Profiles," NBER Working Paper No. 35546 (July 2026)

Why this is riskier than it might seem

Beyond the honesty question, there are two concrete reasons this trend matters for your job search specifically. First, this is now a documented, named pattern that researchers, journalists, and hiring-tech companies are actively studying, which means recruiters are more primed than ever to probe AI claims in interviews rather than take a profile line at face value. A claim that can't survive one follow-up question ("walk me through a specific project where you used that tool") does more damage than no claim at all. Second, the regulatory environment around hiring algorithms is tightening in a way that makes accurate profile data matter more, not less: the EU's AI Act moved recruitment systems into its high-risk category as of August 2026, which requires the data feeding those systems to be accurate and representative. Career-history data, edits and all, is exactly the kind of input those systems draw on.

A better way to show real AI skills

If you've genuinely picked up AI-adjacent tools or workflows, whether that's using an LLM for documentation, working with an AI-assisted analytics platform, or something more technical, there's a more credible way to reflect that than quietly rewriting a job you left years ago:

  • Add it to your current role or a dated skills update, not a closed-out one. If you learned a tool recently, it belongs in your most recent or current position, or in the dedicated Skills section, where the timing makes sense on its face.
  • Be specific about the tool and the outcome, not the buzzword. "Used [specific tool] to cut reporting time by X hours a week" reads as far more credible, and is far more interview-proof, than adding "AI" to a five-year-old job title.
  • Use Licenses & Certifications for anything formal. A completed course or certification has a natural, verifiable date attached to it and doesn't require touching old job history at all.
  • Leave old entries alone unless you're correcting a genuine error. If a past job's description was inaccurate when you wrote it, fixing that is reasonable. Adding capabilities that weren't part of the job at the time is the pattern the research is flagging.

What recruiters and hiring tools are starting to notice

It's not only academics paying attention to this pattern. Recruiting teams that use LinkedIn Recruiter and similar tools can already see how recently a profile section was updated, and a burst of edits to years-old job entries, especially ones that line up suspiciously well with a current job posting's keywords, is the kind of signal that experienced recruiters are trained to notice and probe in a screening call. The researchers behind the NBER paper also point out a second-order effect worth knowing about if you work in data, HR tech, or anywhere near hiring systems: because career-history data feeds labor-market research and increasingly feeds AI hiring tools themselves, a wave of retroactively inflated AI claims doesn't just risk one person's credibility, it quietly distorts the training data those systems learn from. That's part of why regulators are starting to pay closer attention to where hiring-algorithm data actually comes from.

The bottom line

Nearly 1 in 5 LinkedIn users are editing old job history, and AI keywords are the fastest-growing thing being added, closely tracked now by researchers and, increasingly, by the hiring systems recruiters use. The safer and more effective move is to put real, current AI-adjacent skills where they actually belong on your profile, described specifically enough to hold up in an interview. If you want a second set of eyes on how your LinkedIn profile positions your actual skills, that's exactly what our services cover, from headline and summary rewrites to interview prep for the questions a strong profile invites; see pricing for what's included.