Giving feedback is one of the most crucial activities in the workplace. It can also be one of the most difficult. For years, studies have highlighted just how much workers and their leaders loathe performance management, largely because they believe it to be ineffective. A recent Gallup survey found that only 2% of CHROs believe their performance-management system actually works.
The crux of the problem is a lack of trust. In Deloitte’s 2025 Global Human Capital Trends survey, 61% of managers and 72% of workers could not confidently say they trust their organisation’s performance-management process.
But advancements in AI and employees’ appetite for more frequent, real-time feedback have prompted some HR teams to rethink the process. Last week, JPMorgan introduced a new chatbot to assist with the bank’s performance reviews, allowing staff to use it to draft evaluations while keeping pay and bonus decisions at managers’ discretion. Other financial firms, including Citi and IBM, have also deployed in-house AI tools to help write year-end appraisals.
For overstretched managers juggling large teams, the benefits are clear: AI offers a handy shortcut to the often painstaking process of writing multiple reviews. Boston Consulting Group recently reported that AI-assisted drafting of performance reviews can reduce writing time by up to 40%, highlighting some of the potential efficiency gains from using AI.
Convenience comes at a cost
But the latest trend in performance management also shows how increasingly blurred the line between text written by humans and those generated by machines has become.
Managers who chose to partake in the cognitive offloading to machines, without applying their own judgement, risk the insidious consequences of AI-generated workslop: low-quality, mass-produced content, created using AI that looks polished, but lacks real substance. When it comes to evaluating people’s work, the result may be feedback that sounds professional but is filled with non-specific drivel that is unhelpful, incomplete or missing crucial context.
This carries real costs for companies. The most unsurprising being the erosion of trust and respect between managers and their teams. When employees receive feedback that feels generic or formulaic, they are less likely to believe it – or to believe in the person giving it. In fact, a Harvard Business Review study found that 42% employees viewed colleagues who sent poor-quality AI output or “slop” as less trustworthy and 37% saw that colleague as less intelligent.
Furthermore, when it comes to interpersonal dynamics – a critical part of the performance review process – research reveals that most people think chatbots are decidedly useless at giving advice or tackling emotional issues.
Instead of motivating staff, managers who lean too heavily on AI systems to write employee evaluations may unwittingly alienate the very people they are meant to inspire and support. What’s more, they risk doing so at a time when employee engagement is already dangerously low: only 10% of British workers feel motivated in their job, according to Gallup’s 2025 State of the Global Workplace report.
As businesses scramble to address this crisis, automating one of the few remaining opportunities for genuine human connection at work is not the solution. It signals, instead, a troubling abdication of leadership and human oversight at a time when employees are at their most vulnerable.
Moves to outsource the writing of performance reviews to machines is sign that the exercise has devolved into little more than a box-ticking ritual. What was once intended to foster growth, reflection and dialogue between managers and employees has become a bureaucratic compliance task. If managers now need AI to tell them what their employees have achieved over the past year, it reveals how hollow and perfunctory the process has become. Making it even more robotic will only make matters worse.
Giving feedback is one of the most crucial activities in the workplace. It can also be one of the most difficult. For years, studies have highlighted just how much workers and their leaders loathe performance management, largely because they believe it to be ineffective. A recent Gallup survey found that only 2% of CHROs believe their performance-management system actually works.
The crux of the problem is a lack of trust. In Deloitte’s 2025 Global Human Capital Trends survey, 61% of managers and 72% of workers could not confidently say they trust their organisation’s performance-management process.
But advancements in AI and employees’ appetite for more frequent, real-time feedback have prompted some HR teams to rethink the process. Last week, JPMorgan introduced a new chatbot to assist with the bank's performance reviews, allowing staff to use it to draft evaluations while keeping pay and bonus decisions at managers’ discretion. Other financial firms, including Citi and IBM, have also deployed in-house AI tools to help write year-end appraisals.
