AI Tools Create Evidence Gap Between Employees and Managers During Performance Reviews

Employees are increasingly using artificial intelligence to compile detailed evidence for their performance reviews, while managers continue to rely on memory, creating a conflict over evaluation grounds. VentureSquare reported the story.

Some employees now use AI to organize messengers, emails, and work documents into comprehensive performance lists in a matter of minutes. In one instance, an employee submitted a 12-page document as evidence for their evaluation. When the employee contested their rating, the manager cited missed deadlines but could not provide specific dates or evidence, leaving human resources personnel struggling to decide which source to trust.

This shift has highlighted a gap in evidence because employees have extensive digital records of their own work, whereas managers find it difficult to track the yearly activities of multiple team members. Furthermore, AI-generated lists often highlight successes while omitting failures or assistance received from other teams, which may still be remembered by the manager.

Managers have also begun adopting AI for evaluations. A 2025 survey of 1,000 American workers by ZeroBounce found that 41% of managers have used AI to write or edit performance reviews. In the same survey, 26% of employees suspected their reviews were AI-generated. However, AI cannot create records that do not exist, meaning smooth prose does not necessarily equal sufficient evidence.

To resolve these disputes, a three-step process is suggested. First, evidence must be separated into facts, such as submitting a proposal, and interpretations, such as claiming the proposal led to a reduction in processing time. Second, managers and employees should coordinate differing interpretations by identifying which facts led to their respective conclusions or referring back to goals set at the start of the year. Third, the coordinated results, including agreed facts and unresolved disagreements, must be formally recorded.

Organizations are encouraged to integrate these coordination steps into their official evaluation schedules rather than simply having managers comment on submitted evidence. While AI can help categorize facts and interpretations or summarize recorded interviews, the core solution is a structured organizational process for verifying evidence rather than a battle of AI-generated documents.