OpenAI’s latest research explores how people are using ChatGPT. Three big categories dominate the field. Writing (of which, thankfully, only a small portion is dedicated to producing fiction novels); Seeking information; and practical guidance.
In short, people are largely using the system for search, tarting up emails, and, worryingly, for advice and counselling.
The trouble for OpenAI is, none of these are uniquely provided by their solutions. Copilot, integrated into Microsoft Office, seems better placed to help writers. Google will have a thing to say about search. And practical guidance is set to be disrupted by trusted applications from governments, health authorities and a range of other applications that over time will benefit from missing the “big tech ick factor”.
Indeed, research suggests platform usage is tailing off. With even ChatGPT’s most loyal users finding less reasons to engage with the tool.
There is some light at the end of the tunnel. Among work-related conversations, use of the platform changes slightly. Writing still takes up the body. But making decisions and interpreting information climb higher up the list. For the enterprise, which will almost certainly become the key revenue source for the sector, helping workers become more productive is the target. OpenAI’s ability to rise higher up the value chain—from correcting typos in emails to helping make decisions based on complex data—will be the deciding factor in meaningful adoption.
Utility value matters when the technology isn’t cheap. While pricing is increasing, particularly for premium users and businesses, access to tools like ChatGPT continue to be largely funded by private equity. Were users to pick up the total cost today, they probably wouldn’t. OpenAI has two competing forces it needs to work on, bringing costs down, while bringing utility up. Helping the average user conjure up poetry and recipes will achieve neither.
