\n
![](\"https:\/\/gizmodo.com\/app\/uploads\/2025\/09\/anthropic-e1768598845926-1280x852.jpg\")
<\/p>\n
\n<\/p>\n
One of the major promises of AI, whether you buy it or not, is that it will significantly amplify productivity. Who benefits from that productivity depends on who can actually access it.\u00a0AI startup Anthropic is warning that, because of the cost and infrastructure needed to adopt this technology, those potential gains are more likely to make their way to rich countries, exacerbating the already stark levels of economic inequality experienced by lower-income nations.<\/p>\n
The concern stems from a recent analysis conducted by Anthropic to determine how its Claude chatbot is being used around the world. After looking at more than one million conversations from individuals using both the free and paid versions of the chatbot and another million conversations had with the Enterprise version of the service, Anthropic found that richer countries are adopting AI at a significantly faster rate than lower-income countries, and there is \u201cno evidence yet that lower-income countries are catching up.\u201d<\/p>\n
The general thrust of the conclusion is in line with other recent research. Microsoft, for instance, just published a report about AI adoption that found that the \u201cglobal north\u201d nearly doubled the rate of adoption over the last year compared to the \u201cglobal south,\u201d and found that overall adoption is still much higher in richer nations. Peter McCrory, Anthropic\u2019s head of economics, told the Financial Times that if productivity gains materialize, \u201cyou could see a divergence in living standards\u201d that favors areas that are already wealthy.<\/p>\n
Of course, those productivity gains would actually have to happen first. Thus far, it\u2019s not entirely clear whether that has been the case for early adopters. A study from MIT last year found that 95% of businesses that have invested in generative AI tools have yet to produce a net-positive return on investment. Why? Well, ask the workers. According to a study from Upwork, about half of employees surveyed said they don\u2019t know how to achieve the productivity gains their employers expect from AI, and more than three-in-four said AI tools have actually decreased their productivity and added to their workload.<\/p>\n
That\u2019s not to say AI doesn\u2019t have a role to play in boosting productivity or won\u2019t improve over time, just that it\u2019s not quite as simple as \u201cadoption equals gains.\u201d And it\u2019s worth noting that increasing productivity doesn\u2019t mean everyone gets richer. American workers have nearly doubled their productivity over the last half-century, thanks in part to technological advancements. Their wages, though, haven\u2019t kept up, while corporate profits and executive pay skyrocketed during the same period.<\/p>\n
Anthropic\u2019s warning is worth heeding, in the sense that it actually acknowledges that income inequality is a real thing\u2014unlike, say, Elon Musk, who recently suggested that people shouldn\u2019t even bother saving for retirement because we\u2019ll all have universal high income soon. After all, AI will make everything incredibly cheap and accessible. But it is worth asking, if you fear that you may be complicit in building the inequality machine, why keep building it? Anyway, apropos of nothing, Anthropic CEO Dario Amodei has an estimated net worth of $3.7 billion. Not sure why that came to mind.<\/p>\n<\/p><\/div>\n