AI and Productivity: The Impact of ChatGPT's Release on Blogging

Abstract

Artificial intelligence (AI) is one of the fastest-growing technologies in history, offering innovative tools that can enhance worker productivity. In this paper, we examine the influence of ChatGPT's release on the output of workers who adopt it in a quasi-experiment. We study the productivity, popularity, and novelty of bloggers and their work on Medium.com. Popularity and novelty reflect productive efforts to engage one's readers and expand the scope of one's work, respectively. We measure AI use with the least biased detection tool available at the time of this study. We use Coarsened Exact Matching to match ChatGPT-adopters with similar bloggers who do not adopt. We then perform a difference-in-difference analysis around ChatGPT's release using this sample. On average, ChatGPT's release increases the productivity of adopters and increases the novelty of their work but decreases its popularity. Further, the way that ChatGPT is used determines the impact on output. Bloggers with a large number of followers tend to use ChatGPT as a researching tool to create work on new topics, increasing the novelty of their work but at the cost of lower popularity. While tech-savvy bloggers adopt ChatGPT as a writing tool, increasing their productivity but not affecting their work's popularity nor novelty. Finally, the effect on productivity is persistent over the three months following ChatGPT's release.

Keywords: ChatGPT, generative AI, worker output, user-generated content, content creation