Survey reveals AI-generated performance reviews are becoming detectable to employees

A recent survey indicates that employees can frequently identify when their performance reviews are generated by artificial intelligence, suggesting that the use of AI in human resources is becoming transparent to the workforce. According to a report by ZDNET Security, this detectability prompts significant questions regarding whether AI assistance helps or hurts the quality of professional feedback.

The findings highlight a growing tension between the efficiency gains offered by generative AI and the need for genuine, personalized communication in professional settings. As organizations increasingly adopt AI tools to streamline administrative tasks, including the drafting of performance evaluations, the line between human-authored and machine-generated text is becoming less distinct to the recipients. The survey data suggests that this transparency may undermine the intended impact of performance reviews, which rely heavily on perceived sincerity and specific, individualized attention.

Context and Background

The integration of artificial intelligence into workplace routines has expanded rapidly over the past year, with many managers turning to large language models to assist with writing tasks. Performance reviews, which are often time-consuming and require careful phrasing, have become a common use case for these tools. However, the distinct stylistic markers often associated with AI-generated text—such as generic phrasing, lack of specific anecdotal evidence, or overly formal tone—may be giving away the source of the feedback.

According to the ZDNET Security report, the core issue is not just the presence of AI, but the perception of its use. When employees suspect their evaluations are not written by their direct supervisors, it may affect their engagement with the feedback. This dynamic challenges the traditional model of performance management, which depends on a trusted relationship between manager and employee. The survey serves as an early indicator that the "black box" nature of AI assistance is fading, forcing organizations to confront how they disclose and manage the use of these technologies in sensitive interpersonal communications.

Who Is Affected

This development primarily affects knowledge workers and corporate employees in organizations that have adopted generative AI tools for administrative support. Managers who rely on AI to draft initial versions of performance reviews are the immediate actors in this scenario, while their direct reports are the recipients who are identifying the AI involvement. Human resources departments are also impacted, as they must navigate the policy implications of AI-assisted evaluations.

The survey results suggest that this is not an isolated incident but a broader trend within the technology and business sectors. Professionals in these fields are likely to encounter AI-generated content in their daily workflows, making the ability to detect such content a relevant skill. For security and technology professionals, this highlights the importance of understanding how AI adoption in non-technical departments can influence organizational culture and employee trust.

What should you do?

Organizations should evaluate their current policies regarding the use of AI in employee communications. Managers should be trained to ensure that any AI-assisted drafts are thoroughly reviewed and personalized to reflect genuine managerial insight. Employees should remain aware of the potential for AI-generated content in professional feedback and seek clarification if reviews feel generic or impersonal. Companies may need to establish clear guidelines on disclosure, ensuring that employees are informed when AI tools are used in the evaluation process.

Background: Generative AI

Generative AI creates content from prompts. Agentic AI performs actions using tools. The former outputs text or code; the latter changes system states. Understanding this distinction is critical for security architecture, as agentic systems introduce execution risks that pure generation models do not possess.

Read the full guide: Generative AI vs Agentic AI: Architectural Differences and Security Risks

What to do and how to stay safe: artificial intelligence

• Review organizational policies to ensure they address the appropriate use of AI in sensitive communications like performance reviews.
• Train managers to personalize AI-drafted content significantly, ensuring feedback reflects specific, individual contributions rather than generic statements.
• Encourage open dialogue between employees and managers about the tools used in the evaluation process to maintain trust and transparency.
• Monitor for signs of disengagement or confusion among staff that may indicate a lack of clarity regarding the source of performance feedback.

Step-by-step guide: Prompt Injection Warning Signs: How to Spot AI Manipulation

General security guidance from the Threat Times newsroom. It is not confirmed advice from the organisations named in this story.

Frequently asked questions

Can employees really tell when AI writes their performance reviews?

Yes, according to a survey reported by ZDNET Security, employees are increasingly able to identify when performance reviews are generated by artificial intelligence.

Does using AI for performance reviews help or hurt feedback?

The survey raises questions about whether AI use helps or hurts feedback, noting that the detectability of AI-generated text may impact the perceived quality and sincerity of the reviews.

Who is most affected by this trend?

Managers who use AI to draft reviews and their employees, who are the recipients of this feedback, are the primary groups affected by this trend.