LinkedIn's Great Skill Cull: Workers Erase Their Past for AI Futures

It is a peculiar form of professional time travel. Workers across the United States are logging into LinkedIn and systematically rewriting their career histories, inserting mentions of artificial intelligence, GPT, and large language models into job descriptions they held years ago—before ChatGPT existed, before anyone was seriously talking about AI in their work, sometimes before these technologies had names. The National Bureau of Economic Research tracked 29.4 million US LinkedIn profiles and found that nearly a fifth had retroactively changed the title or description of a job the person had already left. The motivation is transparent. Retroactive additions of AI-related terminology to previous roles have risen more than sixfold since ChatGPT's launch in late 2022. A 2026 snapshot of LinkedIn would overstate how common AI-related skills were in 2022 by approximately 30 percent.

This is not harmless personal branding. This is workers engaged in what amounts to systematic historical revision because the labor market has decided that yesterday's experience no longer counts unless it can be repackaged through the lens of today's obsessions. The pattern reveals something darker than resume inflation: it reveals a workforce panicked enough to edit out entire dimensions of their professional identities.

Consider the mechanics. A software developer who spent 2019 to 2022 building scalable backend systems now lists that role as having required "AI integration" and "LLM optimization." A product manager who managed user research and feature prioritization discovers that what they actually did was "AI-driven product strategy." The older skills—the demonstrable expertise that took years to accumulate—do not disappear from the internet. They simply vanish from the version of professional history that algorithms see. LinkedIn profiles are not just personal documents anymore. They are the primary substrate through which hiring algorithms sort candidates, and in a market where an AI-on-AI war already shapes hiring decisions, being invisible to the algorithm is professionally fatal.

The timing is revealing. By the end of 2025, workers were adding remote-work language about as often as they were deleting it. But diversity and inclusion terms fell sharply in early 2025—another inconvenient skill that has fallen out of fashion. Workers are not adding skills to their repertoire. They are performing triage on their own résumés, deciding which aspects of their experience will be legible to the models that evaluate them. The technology boom since ChatGPT's launch has created a retroactive hiring market where what you claim to have done matters more than what you actually did.

Europe has just begun treating this as a compliance question. On 2 August, AI systems used in recruitment and employment decisions moved into the AI Act's high-risk tier, which means employers cannot claim ignorance about algorithmic bias in hiring. Yet the problem is not only on the employer side. Workers are actively collaborating in their own misrepresentation because the alternative—being passed over for jobs because their actual experience is insufficiently trendy—feels worse.

There is a structural damage here that goes beyond individual careers. Career history is the substrate under a great deal of labor-market research. When 29.4 million workers are systematically altering their employment records to appear more AI-competent than they actually were, the data that economists and researchers use to understand workforce skill evolution becomes corrupted. You cannot accurately measure AI adoption in the labor market when workers are retroactively claiming AI expertise they did not possess. You cannot understand what skills actually matter when workers are desperately scrubbing evidence of the skills that got them hired in the first place.

The cruel irony is that this collective act of professional reinvention may be entirely unnecessary. Workers with genuine experience in other domains—distributed systems, user experience, data analysis, project management—have skills that translate to AI work. But the market is not asking for translation. It is asking for the magic words. So workers oblige, editing their pasts into fiction, because in a labor market increasingly sorted by algorithm rather than judgment, fiction is what gets hired.