OpenAI Releases ChatGPT Work Activity Data Across 90+ Countries

Introduction

OpenAI has published a new dataset that tracks how ChatGPT is being used for work across nearly 100 countries. The release, part of OpenAI's ongoing Economic Research initiative, connects real-world ChatGPT usage to a standardized framework of 165 work activities. This gives researchers a detailed view of which tasks are being supported by AI and how adoption varies globally.

The dataset spans from the first quarter of 2025 through the third quarter of 2026 and is broken down quarterly. It covers work-related messages sent through individual ChatGPT plans, excluding enterprise accounts and Codex usage.

By linking ChatGPT activity to Intermediate Work Activities (IWAs), a classification system used in occupational research, OpenAI enables researchers to compare AI usage with traditional employment data. This opens new avenues for studying the economic impact of generative AI.

Key Findings from the Dataset

One of the most notable trends in the data is the growing specialization of ChatGPT usage across countries. In the ten most common work activities, the share of messages declined from 56.0% in Q1 2025 to 51.1% in Q3 2026 across 74 countries observed in every quarter. This suggests that while core use cases remain dominant, countries are increasingly focusing on niche tasks.

The most frequently cited work activities include preparing instructional materials, developing marketing content, programming, and analyzing business data. However, the data also shows substantial variation between countries, indicating that local economic structures and access to AI tools influence how ChatGPT is adopted.

Interestingly, the distribution of ChatGPT usage differs significantly from the broader U.S. job market. About 43% of the employment-weighted task mix for mapped U.S. occupations falls into 'Other IWAs,' which include activities such as moving materials, operating vehicles, and maintaining medical records. These tasks did not meet the threshold for consistent representation in the original ChatGPT sample.

Research Applications and Methodology

The dataset is built on the O*NET framework, a widely used system that groups occupation-specific tasks into broader categories of work activities shared across roles. This standardization allows researchers to compare AI usage with other economic indicators and employment statistics.

Researchers can use the interactive tools provided by OpenAI to explore work activities within specific countries over time, drill down into detailed task categories, and analyze shifts in usage patterns. The data is designed to support deeper inquiry into how AI is reshaping the global workforce.

To protect user privacy, OpenAI does not have researchers read individual conversations. Instead, automated systems classify messages after removing personally identifiable information. The company publishes aggregated statistics, withholds results for small groups, and applies statistical noise to prevent identification of individual users.

What Professionals Should Do Now

For IT, security, and engineering leaders, this dataset highlights the expanding role of AI in everyday work processes. Organizations should take steps to understand how AI tools like ChatGPT are being used internally, especially in areas such as content creation, software development, and data analysis.

• Audit current AI usage: Identify which teams and departments are already using ChatGPT and for what purposes.
• Establish governance policies: Define acceptable use guidelines that align with data protection and compliance requirements.
• Invest in training: Equip employees with the skills to use AI tools effectively and responsibly.
• Monitor for risks: Watch for unintended exposure of sensitive information through AI interactions.
• Engage with research: Use datasets like this to benchmark your organization's AI adoption against industry trends.

Security teams should also be aware that consumer-grade AI tools may introduce new attack vectors, particularly if they are used to process confidential or proprietary data without proper oversight.

Conclusion

OpenAI's release of ChatGPT work activity data provides a rare window into how AI is being integrated into professional workflows across the globe. For researchers and practitioners alike, the dataset offers a foundation for understanding both the opportunities and challenges posed by generative AI in the modern economy. As companies continue to adopt AI at scale, platforms like Grivyonx can help automate compliance and security workflows to ensure safe and efficient deployment.