AI Is Writing Performance Reviews. What Could Go Wrong?

Core Insights - JPMorgan Chase & Co. is allowing managers to use artificial intelligence (AI) for writing performance reviews, which may alleviate the burden of this annual task while raising concerns about the quality of feedback provided to employees [1][2] Group 1: AI Integration in Performance Reviews - The use of AI in annual assessments can save time for managers and potentially yield more useful feedback compared to traditional human-only evaluations [2] - JPMorgan's guidelines emphasize that AI should assist but not replace human judgment, explicitly prohibiting its use for assigning performance scores or making pay and promotion decisions [2] Group 2: Employee Perception and Credibility - Employees may perceive AI-assisted reviews as less credible, questioning whether their managers genuinely believe the feedback provided [5] - There is a risk that AI-generated reviews could be overly positive, failing to address performance issues adequately, which could lead to employee cynicism regarding the review process [5] Group 3: Broader Industry Trends - Companies are increasingly evaluating how employees utilize AI, with firms like KPMG and Shopify incorporating AI usage into their review criteria [6] - Other companies, such as Rippling, are developing systems that assess employee performance more comprehensively, categorizing new hires based on their output during the initial 90 days [8] Group 4: Limitations of AI in Performance Evaluation - HR experts caution that AI has limitations in performance evaluation, emphasizing the need for a clear definition of success at both the company and individual job levels [9][10] - Effective performance reviews still require thoughtful consideration of what constitutes good performance, indicating that AI cannot fully replace human insight [10]