After the launch of ChatGPT, professionals quickly rebranded past projects as "AI-related," similar to how programmers collectively changed "web development" to "mobile experience" during the mobile internet boom of the 2010s. The Stanford team quantified this "time travel" editing behavior for the first time using monthly archival data from Revelio Labs.

Capital and recruitment algorithms reinforce keyword matching, with LinkedIn searches and corporate ATS systems prioritizing profiles tagged with AI, prompting users to actively modify historical records to enhance visibility, motivated by the desire to reduce the risk of being filtered out by algorithms and to compete for limited interview opportunities.

This is highly similar to the collective addition of ".com experience" to resumes after the early 2000s internet bubble and the large-scale addition of "remote collaboration" descriptions during the rise of remote work. We are currently in a phase of skill signal inflation, where real abilities and label density are beginning to disconnect.

Essentially, this is a form of technological substitution. Generative AI has lowered the costs of content production and resume optimization while raising the explicit requirements for "familiarity with AI" on the recruitment side, creating a self-reinforcing cycle: algorithms reward keywords → users retroactively inject keywords → signals become further saturated.

ABAB News · Cognitive Laws

Popular tags will flow back into historical resumes
What algorithms reward will be rewritten into the past
Signal inflation will eventually dilute the premium of real experience