JPMorgan Chase & Co.’s proposal to enable supervisors to use artificial intelligence to help with performance report writing could ease one of the most dreaded annual tasks for managers. Furthermore, it raises the question of whether assessments published by bots would enhance or detract from the process, especially for employees seeking constructive critique.
Executives and management specialists claim that incorporating AI into annual assessments can save managers time and even yield more insightful feedback than employees would receive from human supervisors alone. But they caution that too much outsourcing might turn assessments into AI shit. (In any event, many managers have already created their own criteria, and bot-written reviews are already available, whether or not employers approve of them.)
The boundaries are going to be redrawn with this technology shift, said Benjamin Levick, who is in charge of operations and artificial intelligence at Ramp, a corporate card company. I’m not sure if they’ll experience extreme discomfort or dehumanisation. Naturally, using AI in a very laborious way carries some risk, but I think there is a way to incorporate AI into more managerial processes.
According to JPMorgan’s guidelines, supervisors may use an internal chatbot to assist with their write-ups, but they are reminded that the technology is not a replacement for human judgment. It also forbids them from using AI methods to assign performance ratings or decide on pay or promotions.
According to Levick, time-pressed managers may be failing their direct reports if they don’t use the resources available to them to evaluate employees’ work more thoroughly.
Levick said, I’d feel awful if an AI analyzed everything I wrote and did, generated a grade, and my manager simply reviewed that report at the end. That’s the wrong way to do it, but to be honest, I would also feel bad if my manager had only a few hours to finish this process and they spent all of that time writing a review based on that very small subset after laboriously reading the 2% of things they could find.
Bosses, like everyone else, are susceptible to bias and faulty memory, and they may give undue weight to the recent past while evaluating a longer period of time, according to Peter Cappelli, a management professor at the University of Pennsylvania’s Wharton School and head of its Centre for Human Resources. Cappelli says AI often provides a more objective assessment than a human manager would.
Employees may view an evaluation as less reliable, though, if they are aware that AI was used. Cappelli said, I don’t think it actually came from my boss, so I get a reason to discount this and not pay any attention to it. “I did well, it said.” Does the boss really think that?
Experimental studies have shown that when humans are presented with AI tools, they often rely less on their own judgement and assign analysis to the technology. That can work both ways in a performance review. Employees may feel processed rather than personally evaluated, according to Cappelli, which would make them more suspicious of a review process that many already view as performative. Furthermore, because chatbots are known to be sycophantic, an AI-assisted review may be overly positive and neglect to adequately point out performance issues (one of the many reasons it’s a good idea to examine any AI-written material for fairness, tone, and accuracy).
Businesses will continue to struggle with their choices long after this season of performance reviews is over, according to the pros and cons of AI as well as its novelty. The decisions may be more obvious now that organisations like e-commerce platform Shopify Inc. and consulting firm KPMG LLP specifically assess how often or how well employees use AI as part of their evaluation criteria.
According to a number of HR experts, AI has little use in performance reviews.
Nora Jenkins Townson, the founder of Bright + Early, an external HR department for tech companies, has been approached by a number of startups who claim to have “fixed” performance reviews by lowering bias and easing middle managers’ workloads. For example, by completing a weekly rating card or entering notes from regular one-on-ones into a system, which would produce something like a performance review at the end of the year.
For them to work well, she said, You already have to have a solid definition of what success looks like, both at a company level and for each job. You still need to take some time to think about what a good performance means to you.
