Customer service employees at a Fortune 500 software company who were given access to Generative AI tools became 14 percent more productive on average than those who were not, with the least-skilled personnel receiving the largest benefit from the use of these tools.
Over a year, Stanford University and MIT researchers examined how generative AI technologies affected firm efficiency.
The study is the first to quantify the workplace impact of generative AI tools. GPT-4 scores are 90th on the bar exam, according to previous studies. In small-scale labs, the tech has been tested on workers’ isolated writing assignments.
Erik Brynjolfsson, head of the Digital Economy Lab at Stanford Institute for Human-Centered AI, said some of these earlier tests highlighted the startling potential of large-language models in the workplace. He claimed the tools’ effects are primarily hypothetical until they’re tested.
Also, Read: AutoGPT is making the internet go crazy
In an interview, study co-author Brynjolfsson noted, “Having people use it for over a year in this company, you get a much better sense of how that translates into real-world productivity.” As far as I know, this is the first real-world application.
Brynjolfsson and MIT researchers Danielle Li and Lindsey Raymond examined more than 5,000 customer care representatives, mostly from the Philippines, across important parameters such as how quickly and effectively they solved clients’ problems. Some agents had access to AI technologies trained on many successful customer service discussions, while others did not. The research did not name the enterprise software company for small and medium-sized US enterprises.