Survey Finds Workers Hoard AI Prompts and Hide Usage From Managers to Keep an Edge

The survey covered employees who use AI at work in organizations based in the UK, the US, Canada, Australia, the EU and Latin America. Among those who admitted to hoarding prompts, 44% said they did not want colleagues to gain the same productivity advantage, while 31% said they had invested the time and effort in developing the prompt and did not want to give that advantage away. The finding points to what the survey describes as a proprietary attitude among workplace gatekeepers who hold back their best AI prompts.

Secrecy extends beyond prompts. The survey found 36% of workplace AI users had hidden their AI use from coworkers, and 21% had hidden it from their manager in order to appear more efficient. Separately, 41% said they were worried about colleagues and managers finding out how much time and effort the technology was saving them, and that they would then be expected to reach the same levels of productivity and speed in future.

Ihor Herasymov, co-founder and chief executive of Use.AI, said the results suggest tension is developing in workplaces around AI. "When a useful AI method becomes part of someone's personal edge, sharing it is no longer a neutral act," he said. "Companies may need to think more carefully about how they reward people who turn private know-how into shared capability, because employees will notice quickly when openness helps the organization while weakening their own position."

According to TechRadar's account of the findings, criticism of AI at the point of use has tended to focus on generative AI and its effect on creative industries, while non-creative workplaces have mainly weighed the reliability of results. The survey suggests a different problem: employees treating effective AI workflows as private assets rather than shared practice.

The practical consequence, TechRadar reported, is that companies where this secretive prompt culture takes hold may have to reconsider the metrics they use to judge employee performance, since output measures can reward individuals for knowledge they deliberately keep out of circulation.