The AI Penalty: Workers Punished for Using AI Honestly
Companies are pushing staff to use AI, then handing the credit to the machine and docking the humans. Researchers call this the AI penalty, and it is pushing white-collar workers to hide how much they rely on the tools their bosses told them to adopt.
The Impact on Workers
July 13, 2026 – 3:22 pm
Image by: Canva / PixelsEffect
Workers share stories of being credited for tasks completed with AI assistance, diluting their own contributions and potentially hindering career advancement. For instance:
- Aubrey: Spent a year optimizing a medical manufacturing process using an AI chatbot (Claude). Presenting the work as her own, she compromised between attributing credit to Claude and revealing her role. Her manager still claimed Aubrey built it in a minute with AI, negatively impacting her annual review.
- Deepak: An IT developer at a Fortune 500 firm in India who started openly crediting coding agents he used. His managers assumed all his good work was done by the machine, potentially stalling his promotion.
The AI Penalty in Practice
Many white-collar workers are now caught between bosses demanding more AI and those docking its users. They’ve begun hiding their reliance on AI tools to avoid the penalty.
Christoph Riedl, a professor at Northeastern University, conducted a meta-analysis of 13 studies across various jobs and found a consistent pattern: managers marked work down when staff admitted using chatbots, assuming the machine did most of it—the AI penalty.
Dodging the AI Penalty
Riedl found that the main way to avoid the penalty is to maintain control over the core task and clearly communicate one’s contributions. This becomes increasingly difficult in a job market already anxious about automation and with managers employing blunt measures like token counting.
Token counting, measuring the frequency of prompts to and from chatbots, reveals nothing about AI contributions. Employees have learned to game these systems by asking pointless questions to appear more active. Companies are now restricting such practices.
Finer tools, like coding assistants that co-author code without specifying lines written by humans or their contribution, also mislead. Managers tend to assume the bot was the primary contributor if AI use is disclosed without specific details.
Tools Aimed at Resolving the Issue
Some researchers are developing tools to pinpoint AI contributions more accurately, aiming to mitigate the AI penalty.