ChatGPT vs Claude: How People Really Use AI — And What It Reveals

Artificial intelligence has quickly woven itself into the fabric of daily life, reshaping how people work, learn, and interact. Two of the most prominent AI models, OpenAI’s ChatGPT and Anthropic’s Claude, are now at the center of a fascinating contrast in usage. Recent reports shed light not only on what people are asking these tools to do, but also on what that tells us about broader social and economic shifts.

While OpenAI’s new research paper dissects millions of ChatGPT conversations, Anthropic’s latest Economic Index Report unpacks how Claude is used in both consumer and enterprise settings. The findings reveal not only the different roles these AIs play, but also where adoption is booming — and who is being left behind.

Findings

OpenAI’s study reveals that non-work use now dominates ChatGPT conversations, a major shift from 2024. In June 2024, 47% of ChatGPT chats were work-related, but by June 2025, that figure had dropped to just 27%, despite daily traffic ballooning from 451 million to 2.6 billion messages. The top uses today include tutoring, practical guidance, recipe hunting, product information, and writing tasks such as emails or editing documents. Surprisingly, coding accounts for only 4.2% of activity.

Breaking usage into three categories — Asking, Doing, and Expressing — OpenAI found that 49% of messages are Asking (seeking information), 40% Doing (completing tasks), and 11% Expressing (sharing opinions). At work, the bulk of ChatGPT’s role involves editing and proofreading text, rather than producing fresh content.

Claude, meanwhile, paints a very different picture. Anthropic’s analysis shows automation now surpasses augmentation in Claude’s ecosystem. In consumer and especially enterprise contexts, directive usage — where users simply assign tasks with minimal input — has risen sharply. By early 2025, automation made up 49.1% of usage compared to 47% for augmentation.

Businesses rely heavily on Claude’s coding capabilities: about 44% of Claude’s API traffic is coding-related, compared to 36% on its public Claude.ai platform. Enterprise reliance is particularly stark, with 77% of Claude’s API use showing automation patterns, often for IT-heavy work.

Geographically, ChatGPT boasts 700 million weekly users worldwide, with faster growth seen in low- and middle-income countries. The user base has also diversified: men once made up 80% of users, but by mid-2025, women accounted for about half of the activity. Young people, especially those under 26, dominate usage.

Claude, by contrast, sees its strongest traction in wealthier regions. The US leads in total usage, with California driving 25% of activity. But Israel takes the crown for per capita usage, with an AI Usage Index score of 7 — seven times higher than expected based on population. Washington, DC, also stands out, with people there relying on Claude nearly four times more often than the national average.

Anthropic warns of an emerging inequality: adoption rates mirror patterns of electricity and internet rollout, where wealthier countries advanced more quickly, gaining outsized productivity benefits. Their data shows a 1% GDP increase correlates with a 0.7% increase in AI adoption. Poorer regions often use Claude mostly for coding, while richer regions experiment with broader applications like education and collaboration.

Ultimately, the reports show a clear split: ChatGPT has become the world’s advice and writing assistant, while Claude is positioning itself as an automation and coding powerhouse.

What Undercode Say:

The reports highlight two very different philosophies in AI adoption — human-centered augmentation vs machine-driven automation.

ChatGPT’s dominance in non-work personal use reveals something deeper about the way people integrate AI into daily life. Instead of becoming purely a corporate productivity engine, it has morphed into a digital tutor, guide, and editor, filling roles that are practical, educational, and expressive. This democratizes knowledge access, enabling students, job seekers, and casual learners to use AI as an accessible second brain. However, it also suggests that companies may not be leveraging ChatGPT to its fullest extent for workflow efficiency.

Claude, on the other hand, seems to embody the enterprise-first AI strategy. Its automation-heavy usage reflects a higher degree of trust in handing over end-to-end tasks, especially in coding and IT-heavy sectors. This is a telling indicator of where the labor market might be heading — tasks traditionally requiring junior developers or analysts are increasingly delegated to AI with little human oversight. Businesses are essentially building “AI employees” into their pipelines.

The geographical divide raises urgent concerns. With Claude thriving in wealthier economies and ChatGPT showing broader grassroots adoption, the global AI landscape risks repeating the digital divide of the internet era. If AI-powered productivity becomes concentrated in rich nations, developing countries could fall further behind in innovation and competitiveness.

There’s also an intriguing generational angle: ChatGPT’s heavy youth user base signals a cultural shift. Gen Z and younger millennials are growing up with AI as their default problem-solving tool, much like older generations grew up with Google. This could mean future workplaces won’t ask whether AI should be used — but how seamlessly it can be embedded.

What stands out most is the trust gap. ChatGPT users want to co-create, edit, and learn with the AI, treating it like a collaborative partner. Claude users, especially in corporate settings, are more willing to hand over the keys, treating it like an autonomous worker. These behavioral differences may shape not only the business models of OpenAI and Anthropic but also the future of human-AI interaction itself.

If this trajectory continues, OpenAI may increasingly become the “AI for the masses,” enhancing personal growth and creativity, while Anthropic could become the “AI for businesses,” quietly automating tasks at scale. The danger lies in whether these two paths will widen social and economic divides — or converge into a more balanced ecosystem of human-AI collaboration.

🔍 Fact Checker Results

✅ ChatGPT’s research paper (May 2024–June 2025) confirms non-work usage dominates, with only 27% being work-related.
✅ Claude’s Economic Index Report states automation surpassed augmentation in early 2025, with 49.1% vs 47%.
❌ No evidence yet supports that AI adoption inequality is as severe as electricity’s rollout, though parallels are speculative.

📊 Prediction

By 2027, the gap between ChatGPT and Claude usage will sharpen: ChatGPT will continue to dominate as a personal AI assistant across diverse demographics, while Claude cements its role as an automation-first enterprise tool. This divergence could fuel both cultural and economic divides, pushing OpenAI and Anthropic into different lanes of the AI economy. However, cross-pollination is inevitable — businesses will adopt ChatGPT for communication and education, while individuals will turn to Claude once they see the productivity gains automation provides. The race will be less about capability and more about trust and accessibility.