Artificial intelligence (AI) is changing society as we know it. Tech companies have claimed for years that AI will make our lives easier and make us more productive. People like Elon Musk even believe that humans will no longer need to work, as we will have ‘universal high income.’ However, things may be a bit more complicated in reality. Anthropic’s latest report has warned that the use of AI could widen the gap between wealthy and low-income countries.
Anthropic says low-income countries are falling behind in AI adoption
Anthropic is an AI startup, backed by the likes of Amazon and Google. The company recently analysed the economic reality behind the use of its Claude AI models. In the report, Anthropic has stated that growing global adoption of AI could intensify economic inequality by favouring wealthier countries over lower-income nations. As Anthropic opens its first office in India, the company points to internal research suggesting richer countries are seeing much faster uptake of AI, with “no evidence” that lower-income countries are catching up.
The concern stems from Anthropic’s recent analysis of over one million interactions with its Claude chatbot, both from free and paid users, as well as enterprise clients. The study found a pronounced disparity – nations with higher incomes are adopting AI tools more quickly and widely than their lower-income counterparts. This finding mirrors broader trends in the tech industry, with Microsoft also identifying a “global north” and “global south” divide in AI adoption rates.
Peter McCrory, Anthropic’s head of economics, told Financial Times, “If the productivity gains…materialise in places that have early adoption, you could see a divergence in living standards.” He emphasised the risk of developed countries pulling further ahead as AI-driven productivity grows.
This could make it difficult for low-income countries in the future as wealthier nations such as the US double down on AI.
Low-income countries use free AI versions more
The company’s data shows marked differences in how AI tools are used across economic contexts. Lower-income countries rely more on free access and use Claude mainly for educational purposes, whereas higher-income countries use the chatbot more often for work or personal productivity. According to Anthropic, “Countries differ in their ability to pay for Claude, and [educational] use cases may be better suited to free Claude usage than complex use cases in work areas such as software engineering,” the report said.
Educational background also plays a role, with Anthropic noting that more educated users are able to extract greater productivity benefits from AI by providing more advanced prompts.
More AI in jobs
Anthropic’s research estimates AI could add 1-2 percentage points to annual US labour productivity growth in the coming decade, especially in complex, knowledge-intensive fields. The analysis suggests that about half of jobs could apply AI to at least a quarter of tasks, up from 36 per cent the previous year. Major users of Claude include the US, India, Japan, the UK, and South Korea, while Brazil and Balkan countries show higher enterprise use, and Indonesia stands out in educational applications.
Despite the promise of AI-driven productivity, early evidence on the actual benefits is mixed. A study from MIT found that 95% of businesses investing in generative AI had yet to see a net-positive return.
The comments come days ahead of Anthropic opening its first India office in Bengaluru. India is currently the second largest market for Anthropic in Claude consumption.
