Quick-to-read HR news & insights
From recruiting and retention to company culture and the latest in HR tech, HR Brew delivers up-to-date industry news and tips to help HR pros stay nimble in today’s fast-changing business environment.
Every other day, there seems to be a prediction of how ChatGPT will change peoples’ jobs forever. HR is no different.
While it may feel like science fiction, companies are already deploying some artificial intelligence to draft performance reviews. HR Brew spoke to one CEO about how he’s integrating ChatGPT to aggregate feedback.
The problems with perf reviews. The performance review process can be time-consuming and frequently takes HR departments months to compile and analyze. And the result can be limited, especially when only a manager is providing feedback or when employees are ranked on a subjective number scale. Performance reviews have become even more difficult as remote and hybrid work removed some potentially valuable (literal) employee visibility. Worse, traditional performance reviews are fertile ground for bias and unintentional negative consequences for some employees. HR departments are tasked with analyzing often inconsistent or unreliable employee feedback, and producing a coherent report that often impacts the employee’s job.
Enter ChatGPT, the language model trained by OpenAI. The emerging chat technology will certainly replace some tasks (especially the super-boring ones), but the AI chatbot won’t take over an HR manager’s job just yet.
One company, Confirm, launched a ChatGPT-powered performance review product last month. HR Brew spoke to Joshua Merrill, Confirm’s CEO, about how its system asks pointed questions to multiple direct-contact colleagues. He says the questions vary but they include, “Who do you go to for help and advice? Who do you believe is an outstanding contributor?”
It then analyzes the data and uses ChatGPT to draft a performance review based on that feedback. “If all you were to do is answer those questions about your colleagues, we could actually predict what the manager would have rated them…about 80% of the time,” Merrill explained.
However, Merrill stressed that while Confirm’s solution is proving fairly effective at writing draft reviews so far, it has limitations because “the technology isn’t at the level where it should be giving an employee feedback—that really has to come from the manager.”
The future. Merrill believes that while there are some obvious shortcomings of ChatGPT right now, the technology will likely look very different in five years. He hopes that leaders will understand “that the way that we work is in networks that are dynamic and diverse and creative. We have many stakeholders to please, not just one manager, but that AI will be playing a role in pulling the insights out of that data and making recommendations based on that data that are accurate.”—KP