Managers devalue contributions when AI is credited, creating a 'AI penalty' that stalls raises and career growth. Researchers find token tracking doesn't capture creative input.
When Aubrey finished a major project earlier this year, her manager asked her to credit Claude, the AI chatbot, for the work. Aubrey, a New York healthcare analyst, had spent over a year developing a new manufacturing process. She used Claude in a small capacity. Her manager wanted the presentation to make it sound as if the AI had conceived and executed the idea on its own.
"I had worked for over a year to gather issues, draft alternatives, learn the implications of any changes, and my manager wanted to credit all of that to AI," said Aubrey, who requested anonymity for fear of retaliation.
She compromised by inflating AI's role while still communicating her own effort. During the presentation, her manager interrupted and announced that she had built it all out in a minute with AI. Weeks later, Aubrey received a lukewarm annual review. Her boss later said the incident was a factor.
Deepak, an India-based IT developer for a Fortune 500 tech company, faced a similar problem. He started crediting the automated coding agents he uses for grunt work. Upper management began assuming all his positive contributions came from AI. He suspects it stalled a promotion.
Christoph Riedl, an information management professor at Northeastern University, said the hesitation to disclose AI use is justified. He and his coauthors examined 13 studies covering a range of job functions. The finding was consistent: managers devalued workers' contributions when AI assistance was revealed.
"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 said.
Companies are adopting methods to track AI use. Most rely on token monitoring, which measures how often employees query the chatbot and the volume of information exchanged. These metrics do not capture creative input. Anyone could ask endless irrelevant questions and still appear to be a pro AI user.
Amazon recently shut down an internal leaderboard that tracked AI token use after it encouraged frivolous queries. Dave Treadwell, an Amazon senior vice president, told staff at a companywide meeting: "Please don't use AI just for the sake of using AI."
Some researchers are developing better attribution tools. Graham Neubig, a computer science professor at Carnegie Mellon University, cofounded OpenHands, an open-source platform that adds footnotes to AI-generated code. Neubig said it was important to tag the code to moderate trust and increase scrutiny.
IBM created the AI Attribution Toolkit, inspired by the Contributor Role Taxonomy used in scientific publishing. The tool lets people specify how much of the work was auto-generated, whether the chatbot produced the content from scratch, and whether certain elements were human-reviewed. The toolkit then produces an attribution statement. Jessica He, one of the designers, said high-level acknowledgments are insufficient.
"A user may feel that attribution encroaches on their ownership if their AI use was limited," she said.
Oliver Schilke, a management and sociology professor at the University of Arizona, said the tension between firms' urge to adopt AI for efficiency and the social costs of its adoption is a central contradiction. His research found that the simple act of disclosure can make people trust you less.
"Those who do the morally right thing must bear the penalty for transparency," Schilke said. He advocates for collective AI governance norms that include tools like the Attribution Toolkit.
Thomas Prommer, an engineering executive at Adidas, saw mandatory attribution kill initiative. Engineers stopped using AI tools because they didn't want their best contributions footnoted as co-written by Claude. "The signal it sent was: AI help diminishes your work. So people hid it or avoided it," Prommer said. What worked was crediting outcomes rather than tools.
Alessio Artuffo, CEO of Docebo, a learning platform, said the question is not how the work was produced but whether the person can defend it and be held accountable when it fails. "If employees are producing more output but feeling less ownership of the work, that's not a win; that's capability regression dressed up as efficiency," he said.
The burden falls on individual workers to decide how much to involve AI and how much to reveal. Schilke said the dynamic creates a paradox: those who are transparent about AI use pay a price. Researchers and organizational experts agree that if companies want employees to use AI creatively, they need to build environments where AI proficiency is valued rather than penalized.
IBM's IBM stock page Alpha Score of 61/100 reflects a moderate outlook. The company's AI attribution toolkit shows how difficult it is to credit human versus machine work. For a broader look at how technology is reshaping workplace dynamics, see stock market analysis.
Drafted by a large language model from the source reporting linked above, then screened by automated publishing checks. It is not read by a journalist before publication. Some articles cite our Alpha Score. Verify prices and figures against the original source. Educational coverage, not personalized advice.