Two landmark reports — one from OpenAI and Harvard economist David Deming, and another from Anthropic’s Economic Research team — provide the most detailed look yet at how AI tools are being used across work and personal life.

The studies reveal both commonalities and contrasts: while ChatGPT is largely used for practical advice, writing, and decision support, Claude shows a sharp shift toward automation, especially among business and API users. Together, the findings underscore how consumer adoption is broadening globally and how AI is beginning to reshape economic activity in measurable ways.

OpenAI and Harvard economist David Deming have released the largest analysis of ChatGPT consumer usage, published as a working paper through the National Bureau of Economic Research (NBER). The study examined 1.5 million conversations, offering the most detailed view yet into how the AI system is used across work and personal contexts.

The report highlights how adoption has broadened since ChatGPT’s launch in 2022. By mid-2025, the platform reached 700 million weekly active users on consumer plans, underscoring its role as one of the most widely adopted AI technologies globally.

“Access to AI should be treated as a basic right—a technology that people can access to unlock their potential and shape their own future,” OpenAI stated in the blog post.

The study shows that early demographic gaps are narrowing. In January 2024, just 37% of users with gendered names had typically feminine names. By July 2025, that figure had risen to 52%, signaling more balanced adoption between men and women.

Geographic access has also broadened. Growth in low- and middle-income countries outpaced higher-income nations. By May 2025, ChatGPT adoption in the lowest-income countries grew at more than 4x the rate of the highest-income countries.

Most consumer usage centers on everyday tasks. Roughly 75% of conversations involve practical guidance, information seeking, or writing, with writing being the most common work task.

Researchers categorize usage into three types:

ChatGPT’s economic impact spans both professional and personal contexts. Approximately 30% of consumer usage is work-related, while 70% is non-work, with both categories continuing to expand over time.

In many cases, the value created goes beyond what traditional measures like GDP can capture, reflecting how people use ChatGPT not only for productivity at work but also for personal benefit. A key source of this value is decision support: the system helps users improve judgment and efficiency, especially in knowledge-intensive jobs.

As adoption broadens, user groups increase activity over time as models improve and new use cases emerge, underscoring ChatGPT’s dual role as both a workplace tool and a driver of value in everyday life.

For full results and methodology details, see the complete working paper.

Q: What is the ChatGPT usage study?
A: It is a National Bureau of Economic Research working paper analyzing 1.5 million conversations to track how consumer adoption has evolved.

Q: How many people use ChatGPT weekly?
A: Around 700 million weekly active users send 18 billion messages through consumer plans.

Q: How has gender adoption changed?
A: The share of users with typically feminine names rose from 37% in early 2024 to 52% in mid-2025, showing a narrowing gender gap.

Q: What are the main use cases?
A: About 75% of conversations focus on practical guidance, information seeking, and writing, with Asking (49%), Doing (40%), and Expressing (11%) as key categories.

Q: How much use is work-related?
A: Approximately 30% of conversations are tied to work tasks, while 70% are for personal or non-work purposes.

Anthropic released its third Economic Index Report, a large-scale analysis of how people and businesses use Claude across countries, U.S. states, occupations, and time. The study uses a privacy-preserving classification method to categorize anonymized interactions into task groups defined by the U.S. government’s O*NET database.

In this latest report, Anthropic measures how Claude is being used differently across four dimensions:

The findings highlight both uneven adoption and a shift toward automation. Claude is being used heavily for software development, but tasks in education, science, and creative work are expanding rapidly. Businesses, in particular, are showing a strong tendency to let Claude complete tasks independently.

Claude usage is dominated by the United States, followed by India, Brazil, Japan, and South Korea. But when adjusted for population size using the Anthropic AI Usage Index (AUI), smaller high-income countries like Israel and Singapore emerge as leaders.

Anthropic found a strong link between GDP per capita and Claude adoption: a 1% rise in GDP per capita correlates with a 0.7% higher AUI. This suggests that countries with stronger digital infrastructure and knowledge-based economies are outpacing others in AI adoption.

The report raises concerns that AI could mirror past general-purpose technologies like electrification, which widened global economic divides if benefits were concentrated in wealthier countries.

Within the U.S., Claude adoption increases with state income, but the relationship is not straightforward. The study found that a 1% rise in a state’s per capita GDP is linked to a 1.8% higher AUI (Anthropic AI Usage Index), meaning that wealthier states generally use Claude more often.

However, income alone does not fully explain these differences. Adoption patterns vary widely depending on the industries that dominate each state’s economy:

These findings show that while income matters, local economic structure is often the stronger driver of how Claude is used.

Since December 2024, Claude has been most used for computer and mathematical tasks (37–40% of conversations). However, knowledge-intensive areas are rising:

At the same time, interaction patterns are shifting. Directive automation rose from 27% to 39%, making automation (49.1%) more common than augmentation (47%) overall. In high-use countries, augmentation is still more common, but in lower-use nations, automation dominates.

For the first time, Anthropic included data from its API customers, allowing a direct comparison between business use and consumer use on Claude.ai. These API users — typically companies and developers — interact differently because they pay per token rather than through a monthly subscription and often integrate Claude into their own workflows.

Key differences stand out:

Anthropic has also released an interactive website where readers can explore Claude.ai usage across countries, states, and occupations, along with the full underlying dataset. For complete results, methodology, and analysis, the full report is also available.

Q: What is the Anthropic Economic Index?
A: A recurring report analyzing how people and businesses use Claude, across geographies, tasks, and time.

Q: How is Claude use shifting globally?
A: Higher-income countries lead adoption, while lower-use countries automate tasks more heavily.

Q: Which U.S. state has the highest usage index?
A: The District of Columbia, with an AUI of 3.82.

Q: What tasks are growing the fastest?
A: Educational instruction (up from 9% to 13%) and science tasks (up from 6% to 8%).

Q: How do businesses use Claude differently?
A: 77% of API traffic shows automation, far higher than the consumer platform Claude.ai.

The two studies highlight that AI adoption is expanding across demographics, geographies, and industries, but in uneven ways. ChatGPT demonstrates how AI can act as an everyday assistant and advisor, while Claude reflects a growing reliance on automation at scale, particularly in business contexts.

Both reports suggest that as models improve, usage deepens and trust in AI systems grows — raising opportunities for productivity gains and new forms of economic value. At the same time, they also point to emerging challenges, including the risk of widening economic divides if adoption remains concentrated in higher-income countries.

What makes these findings significant is their scale and rigor: together, they represent millions of real interactions and some of the first systematic evidence of how AI is reshaping work and personal life. Rather than speculation, these results provide hard data that policymakers, businesses, and researchers can use to prepare for the accelerating role of AI in the global economy.

The shared message is clear: AI is no longer a niche experiment — it is becoming central to daily life and economic activity worldwide.

Editor’s Note: This article was created by Alicia Shapiro, CMO of AiNews.com, with writing, image, and idea-generation support from ChatGPT, an AI assistant. However, the final perspective and editorial choices are solely Alicia Shapiro’s. Special thanks to ChatGPT for assistance with research and editorial support in crafting this article.