LinkedIn Users Rewrite Past Jobs for AI Buzz

A recent NBER paper, "Time Travel on Professional Profiles," explores how LinkedIn users retroactively alter past job titles and descriptions. This "time travel" method reveals significant shifts in the labor market as users, particularly "established" individuals with over 100 connections, have retrospectively emphasized artificial intelligence usage and de-emphasized diversity, equity, and inclusion (DEI) in jobs held between 2020 and 2026. Researchers estimate that at least one in five LinkedIn users has engaged in this practice, often editing descriptions for jobs held years prior. The study indicates that AI product managers are the most represented group making these edits, while dentists are among the least. The paper also notes a rise in users employing AI tools to revise their old job descriptions, with usage declining as education levels increase, although MBAs from less prestigious programs show a notable exception.

These retroactive edits are strongly linked to labor market transitions, with workers more likely to change employers following such modifications. This mutability in profiles can skew historical skill assessments but also offers insights into workers' perceptions of in-demand skills. The trend provides a "supply side" view, contrasting with "demand side" analyses of changing language in job descriptions. For example, the modeling of AI mentions in old job descriptions shows users are increasingly highlighting AI usage now compared to the time they actually held the roles. This behavior is often observed when users switch companies, as internal job changes may involve less scrutiny of how a role is described.