Does adopting artificial intelligence (AI) lift every employee's performance

or does it help some while holding others back? Using 582 frontline employees and a structural equation model with bootstrap tests

we show that AI adoption works through two opposing channels. It raises performance by lowering professional barriers and strengthening perceived skill compensation

the enabling path; and it lowers performance by masking the distinctiveness of an employee's expertise and weakening intrinsic motivation

the masking path. The enabling indirect effect is positive (0.130) and the masking indirect effect is negative (-0.040)

and both are significant

so AI's effect is genuinely double-edged. Which channel dominates turns on the employee's own skill level. For low-skill employees the total indirect effect is clearly positive (0.220)

whereas for high-skill employees it falls to near zero and even turns negative (-0.039). Put plainly

AI narrows the skill gap by lifting weaker performers

yet it can quietly sap the drive of a firm's most skilled people. The study introduces and validates the construct of professional distinctiveness masking and explains why AI is best deployed in a differentiated way.