OpenAI’s new economic analysis
Captured source
source ↗OpenAI’s new economic analysis | OpenAI
July 22, 2025
OpenAI’s new economic analysis
Analysis provides insights into ChatGPT’s impact on the economy. OpenAI also launches new research collaboration to study AI’s broader effects on the labor market and productivity
Loading…
Share
Over half a billion people around the world actively use OpenAI’s AI tools, especially our freely available ChatGPT. They send more than 2.5 billion messages to the platform per day – including more than 330 million per day in the US. This breadth of use offers a unique window onto AI’s impacts on the economy. ChatGPT has saved teachers nearly six hours per week on tasks; saved state workers in Pennsylvania an average of 95 minutes a day on rote tasks so they can deliver better services; and enabled entrepreneurs to build new companies and start-ups.
Today, OpenAI is releasing our first look at how our tools like ChatGPT have boosted productivity for workers across businesses and governments – and how that’s already creating value. The note is authored by OpenAI Chief Economist Ronnie Chatterji and the OpenAI Economic Research team.
We’re also announcing today that Chatterji, along with Jason Furman of Harvard University and Michael Strain of the American Enterprise Institute and Georgetown University, are undertaking a 12-month collaboration to assess AI’s impact on productivity and the workforce. Together, they will develop a research agenda and metrics to analyze AI’s effect on jobs, an effort that will be housed in the new OpenAI Workshop in Washington, DC, along with trainings, demos, and other programming for key economic stakeholders.
ChatGPT is the fastest-adopted consumer technology in history, reaching 1 million users in five days, 100 million users in two months, and over 500 million users today. Since launch, we have seen a dramatic increase in use of ChatGPT for work. Today, 28% of employed US adults who have ever used ChatGPT report using ChatGPT at work, compared to just 8% in 2023.
Throughout history, certain technologies have fundamentally changed what people can achieve – the wheel, the engine, electricity, the transistor. AI is a similar transformative technology. At its core, AI is helping people scale their ability to think, learn, create, and build. It’s scaling human ingenuity itself. As these systems improve, the economic benefits are expected to be significant.
Economists differ in their projections for how AI will impact productivity, but even at the lower end, AI will expand the economic pie. The most pressing questions are: how will that expansion unfold, and who gets what slice?
The choices we make today will shape whether this transformation leads to greater opportunity for all, or greater concentration of wealth and power for the few. At OpenAI, we want everyone to be on the “up elevator” of AI.
We’re already seeing glimpses: scientists accelerating discovery, small business owners expanding their capacity, teachers preparing lessons faster. Imagine what’s possible when these tools are in everyone’s hands.
This extraordinary potential doesn’t mean progress will move in a straight line. Every major technological shift brings change; some jobs disappear, others evolve, new jobs emerge. The pace and scale of change with AI will likely be bigger and faster than what’s come before. We need to be honest about this fact and prepare now.
Our task isn’t to stop this change, but to guide it. History shows that new technologies can expand opportunity and make life better for everyone – but only if we democratize access, support workers through changing times, and build economic systems that reward broad contribution and participation.
You can read our first look at ChatGPT‑driven productivity here.
Author
OpenAI
Keep reading
OpenAI and Malta partner to bring ChatGPT Plus to all citizensGlobal AffairsMay 16, 2026
How ChatGPT learns about the world while protecting privacy Global AffairsMay 6, 2026
How frontier firms are pulling aheadCompanyMay 6, 2026
Notability
notability 5.0/10Substantive post from major lab, not a model release or code