When Will Average People Feel AI’s Impact?

Many AI optimists compare the current boom with the Industrial Revolution or other periods of rapid technological change. These comparisons capture the scale of the transformation, but they overlook how most people experience technological progress. Unlike earlier industrial eras, artificial intelligence has not yet produced many tangible new goods that reshape everyday life. Society also has more inertia resisting change than it did during previous periods of rapid technological diffusion.

AI is still a rounding error in everyday life

For most people, contact with AI products remains marginal, confusing, or only modestly useful. Some associate AI with image generation, enhanced Google Search, or other small conveniences. Others connect it with addictive social-media algorithms, people becoming dependent on chatbots, or debates over data centers.

Core parts of daily life, including family, food, transportation, and entertainment, have experienced few direct effects so far. Outside the technology sector, it is still possible to spend weeks barely thinking about AI. During a recent break, the only uses of AI involved search and creative work, including preparing a seating chart for wedding guests.

Earlier industrial revolutions produced visibly life-changing benefits for ordinary people. The First Industrial Revolution, beginning in the late 18th century, brought cheaper clothing, cooking ware, reading material, and new forms of employment. The Second Industrial Revolution, in the late 19th century, introduced household machines such as sewing machines, preserved food, indoor plumbing, photography, improved lighting, bicycles, and a broader range of manufactured goods enabled by electrification.

Most of these technologies remain part of everyday life. Their benefits were physical, direct, and easy to recognize.

Even the most optimistic forecasts for AI involve new scientific discoveries, advanced treatments for rare diseases, and potentially sustained economic abundance. However, these benefits may be too indirect for people to associate them clearly with AI companies.

If a family doctor recommends a new treatment that saves a patient’s life, how likely is that patient to credit OpenAI or Anthropic for developing it? How many Americans will care about OpenAI solving the Navier-Stokes Millennium Prize Problem?

Fifty years from now, the average American’s daily life may still look broadly familiar. Homes, appliances, relationships, and vehicles could remain similar, although self-driving technology will likely continue to spread. That technology has developed along a trajectory largely independent of large language model innovation.

AI may receive substantial credit for changes that emerge over this period. Fifty years is a long time, particularly given the speed of progress in the narrow portion of technology currently receiving the most attention.

The most important early work of the AI revolution is building foundational infrastructure and a general process that can compound over decades. A major mathematical breakthrough made today may eventually appear minor compared with what follows from that accumulated progress. It remains difficult to predict how much faster the technologies people use every day could improve as a result.

Breaking social stasis

Much of the public narrative around AI appears designed to persuade people to care about its long-term potential. Turning that potential into broad, visible benefits will take a long time, while the industry already faces political pressure because of the imbalance between those benefiting from AI and everyone else.

Today, AI primarily serves the economic elite. In knowledge work, which represents roughly half of the U.S. economy, AI is becoming as fundamental as electricity, or may become so as agent capabilities improve. A tool with such transformative productive power benefiting only half of society creates instability. Many people can see the contrast between a booming technology sector and a broader quality of life that appears stagnant.

This dynamic resembles Engels’ pause, the period from 1790 to 1840 when British working-class wages stagnated while per-capita gross domestic product expanded rapidly during a period of technological upheaval. If the AI era follows a similar pattern, people who do not benefit from the technology will have legitimate reasons to resist it.

AI is highly effective for scaling technology companies and creating online-native small businesses. At the same time, the technology sector may not expand its workforce in proportion to its success. Knowledge-work output could rise sharply even as headcount declines. Because technology companies were already among the most successful parts of the American economy, AI risks being viewed as another tool that benefits a narrow group rather than as a collective good.

That perception could restrict AI’s development and lead to a path resembling the cautionary history of American nuclear power.

The industry also faces intense expectations. Millions of people are watching AI, and the public may not be willing to wait decades for its benefits to diffuse. If AI had a century to spread through society, its effects would likely become much more obvious, as happened with earlier industrial revolutions. Instead, the industry is being judged during what may be only the first half-decade of a 50-year diffusion process.

Two problems are especially important:

1. AI’s positive effects are still too indirect during its early development.
2. AI faces a political backlash tied closely to the history of Big Tech in Western society. This is partly a matter of timing. If AI’s exponential growth had arrived decades after the current platforms had worked through their major problems, data centers might not have become such a central political issue.

Solving either problem would reduce pressure on the industry and give it more time to demonstrate why society should accept changes to the economic status quo. Both challenges are intensified by AI’s own public framing, which often emphasizes danger, catastrophe, and mass unemployment. Industry leaders have begun addressing this problem, but the public has not fully accepted the broader trajectory.

The importance of tangible benefits

Robotics and self-driving vehicles may eventually become closely associated with the current AI revolution. If mass-produced large language models help accelerate robotics and bring capable machines into everyday life, people may quickly recognize the tangible benefits of AI.

There is some irony in this possibility. Many observers have emphasized that current large language models differ from the general AI progress of the previous decade or two. If advances in that broader AI story later make the benefits of language models more visible, the distinction may matter less to the public.

The current period can be understood as AI’s growing pains. Society is working through problems that predate ChatGPT, releasing accumulated frustration and resistance before it can take advantage of longer-term growth. The process of diffusion will likely take much longer than the initial political and social opposition.

People who are young today may see powerful AI progress from effectively 0% adoption to more than 90% adoption during their lifetimes. AI that is deeply integrated into businesses, acts as a personal assistant, and supports ordinary activities is only beginning to become viable. Such systems will take longer to spread than easier-to-understand applications such as ChatGPT, but they may represent the more meaningful measure of AI’s evolution.

This perspective makes continued technological progress important. The eventual benefits could be substantial, but they are not guaranteed, and considerable work will be required to ensure that those benefits are distributed widely.