Managers Devalue Work When Employees Disclose AI Use, Studies Show

The AI attribution dilemma

Aubrey, a healthcare analyst in New York, spent over a year developing a faster medical manufacturing process. She used the Claude chatbot minimally, but her manager insisted she credit the AI as if it had conceived and executed the entire project. When Aubrey resisted during her presentation, her manager interrupted to claim the work was built "in a minute with AI." Weeks later, Aubrey received a disappointing annual review.

Across the globe, Deepak, an IT developer at a Fortune 500 tech company, began transparently crediting the automated coding agents he deployed for routine tasks. Upper management soon assumed all his contributions came from AI. He believes this perception has stalled an expected promotion.

These cases, first reported by Business Insider, illustrate a troubling pattern emerging in white-collar workplaces: employees who disclose AI assistance are penalized for their transparency.

Research confirms the penalty

Christoph Riedl, an information management professor at Northeastern University, led a meta-analysis examining 13 studies across various job functions. The findings were unambiguous: managers consistently devalued workers' contributions when employees revealed AI had assisted them. Managers defaulted to assuming the technology performed most of the work.

"If AI use is disclosed without specific details about how it was used, the manager's default assumption seems to be that it was used in a way that reduces agency," Riedl explained. The few workers who avoided this "AI penalty" retained clear agency over their core work and explicitly outlined their contributions.

Separate research by Oliver Schilke, a management and sociology professor at the University of Arizona, found that AI disclosure can erode trust among colleagues, who may perceive users as lazy—even when the disclosure is made in good faith.

The tracking problem

Most companies now track AI usage through tokens, the fundamental data units processed by AI models. This metric shows query frequency and interaction length but reveals nothing about creative contribution. The system encourages gaming: employees can rack up tokens with irrelevant queries while appearing to be power users.

Amazon recently shut down an internal leaderboard tracking AI token use after recognizing it pushed staff toward unproductive "tokenmaxxing." At a company meeting, senior vice president Dave Treadwell told staff: "Please don't use AI just for the sake of using AI."

Even sophisticated attribution methods fall short. AI coding assistants like Claude Code automatically add co-authorship signatures without specifying which lines were auto-generated or the extent of human involvement.

Emerging solutions

Researchers are developing better attribution systems. Graham Neubig at Carnegie Mellon University cofounded OpenHands, which adds footnote-like attribution to AI-generated code lines. IBM created the AI Attribution Toolkit, inspired by scientific publishing standards, allowing users to specify how much work was auto-generated and whether elements were human-reviewed.

Jessica He, one of the toolkit's designers, notes that high-level AI acknowledgments are insufficient. How people engage with work differs based on whether AI generated new ideas or simply refined wording.

Why it matters

This creates a paradox at the heart of AI adoption: companies push employees to use AI tools for efficiency, but workers who comply transparently face career penalties. The burden falls on individual employees to navigate disclosure decisions alone, while those who do "the morally right thing must bear the penalty for transparency," according to Schilke. Without collective governance norms and better attribution tools, organizations risk either widespread AI concealment or capability regression disguised as efficiency gains. Workers increasingly wonder whether they'll receive credit for AI-assisted successes while taking full blame for AI-generated failures—a concern validated by reports of Amazon laying off employees for mistakes made by AI agents.

Business Insider first reported these findings and worker experiences.