Research suggests AI fear may outpace its near-term economic impact

In short: New analysis suggests AI may boost the economy slowly over the next decade, while fear and mistrust of AI are already having faster effects.

What's going on

Several economists and policy groups say the best evidence points to a modest, uncertain economic lift from AI in the next 5 to 10 years. Economist Daron Acemoglu, in an MIT-linked analysis, estimates that only about 5 percent of tasks in the US economy are likely to be profitably automated by AI over the next decade. He projects productivity gains of around 0.7 percent over 10 years, which could translate to roughly 1.1 percent added to US GDP.

Another study, from Analysis Group, puts AI’s global impact over about a decade in a wide range, from $1.49 trillion to $2.95 trillion. That equals about 0.2 percent to 1.0 percent of global GDP, and the authors stress that no single “right number” is possible. The US Congressional Budget Office also says AI could change how goods and services are delivered, but it emphasizes that timing and size are hard to predict.

At the same time, the analysis argues that public fear is already shaping real-world decisions. Some people worry about job losses, fraud, and deepfakes (fake photos, audio, or video that look real). That fear can affect spending, hiring, and investment, a bit like how rumors about a bank can cause a bank run even if the bank is fine.

What to watch

A key risk is that markets swing too far in either direction. If investment in AI is driven by a stock market bubble, a sudden drop in confidence could hurt growth. On the other side, rushed or overly broad rules could slow helpful uses of AI, while well-targeted rules could build trust and support steadier adoption.

Source: NYTimes