Since generative AI went mainstream a year ago, it has inspired both hype and fear. Advocates predict tools like ChatGPT and DALL-E will transform the economy, while skeptics worry about inaccuracies, harmful results, and the potential for job displacement. However, a recent study by the National Bureau of Economic Research reveals that deploying generative AI at scale can significantly boost productivity without replacing workers.
The study focused on nearly 5,200 customer support agents at a Fortune 500 software firm who received a generative AI-based assistant in a phased rollout from November 2020 to February 2021. This AI tool provided real-time recommendations, suggested responses, and supplied links to internal documents during support chats. The results were impressive: agents with AI assistance were 14% more productive, resolving more issues per hour and handling more chats. The least skilled and least experienced workers saw the most significant productivity gains, up to 35%.
Erik Brynjolfsson, a senior fellow at Stanford’s Institute for Human-Centered Artificial Intelligence, noted that the productivity increase was rapid, without the usual initial decline seen with new technologies. The AI assistant learned from millions of service interactions, disseminating effective practices across the workforce. This helped newer agents perform at the level of more experienced ones.
Customer satisfaction also improved, as reflected in surveys and conversation analyses. Customers appreciated faster problem resolution, and agents were less likely to quit, possibly because the AI made their jobs more enjoyable by improving customer interactions.
Overall, the study concluded that generative AI benefits employees, customers, and stockholders alike. Brynjolfsson believes this technology will significantly boost U.S. productivity over the next decade, potentially raising living standards and reducing inequality. He emphasizes that employers should use AI to augment rather than replace workers, ensuring high-skilled employees are retained to help the system continue learning and improving.
