Technological progress in many ways feels routine in the 21st century. We see it occurring around us everywhere from the constant improvement in solar and other renewable technologies, to the sudden surge of electric vehicles on the road, and even to renewed interest in space travel. But nevertheless, certain technologies stand out because of their broad impacts on society. These sweeping, revolutionary technologies are known to economists as general-purpose technologies, or GPTs.

GPTs are technological innovations usually having effects across many industries and sectors. These technologies have the potential to drive significant economic growth because they enable new forms of production and increase productivity in existing sectors. Policymakers interested in boosting economic growth must therefore have as a major source of focus the discovery and adoption of more GPTs.

Technology and Economic Growth

Nowadays, most economists agree that technological innovation is a critical driver of economic progress. But not long ago, economists thought a nation’s level of savings and its rate of capital accumulation were more important keys to prosperity. This notion fit with the conventional folk wisdom that “a penny saved is a penny earned.” In the earliest economic growth models, a country’s savings rate determined its growth rate.

A refinement of these early models came in the mid-20th century with the now-famous “Solow model,” named after Nobel laureate Robert Solow. The Solow model predicted that countries with lower initial levels of capital would grow more quickly. Due to diminishing returns—i.e., the fact that eventually all the various uses of capital get exhausted—a country’s growth naturally slows as its capital base expands, so richer countries grow more slowly and poorer countries eventually catch up in terms of wealth and prosperity.

The Solow model helped explain the rapid economic growth seen in countries such as Germany and Japan following World War II. Despite experiencing devastating destruction, both countries grew rapidly as they rebuilt their economies. However, there were problems with the early growth models. Many poor countries’ economies seemed to never take off, while rich countries like the United States grew rapidly in the postwar era. Moreover, a lot of countries had higher savings rates than the U.S., yet they grew more slowly.

With improvements in statistical methods and as better data became available, it became possible to untangle some of the determinants of growth. Quite embarrassingly for the Solow model, most of the divergences in growth observed across countries seemed to be explained by the so-called residual in the model, or what was left over that couldn’t be explained by other variables in the model, such as labor and capital.

The lesson learned from this period is that, while savings and capital accumulation do explain a nontrivial component of growth, technological advancements and strong institutions seem to be more important. Moreover, technological advancements in and of themselves are not enough to lead to GDP improvements, as they need to be translated into marketable products. It wasn’t until books were printed and sold in the marketplace that the full potential of the printing press was realized. As a consequence, the beneficiaries of innovations are not always the same as the people doing the inventing. This translates into a misalignment of creation and rewards. Correspondingly, there is not enough incentive to innovate relative to what would be ideal.

The Mixed Blessing of Innovation

Perhaps the first GPT that man stumbled upon was fire. A tell-tale sign that fire is a GPT is its many uses. It can provide warmth, cook food, clear away unwanted forest and foliage and forge metal. It played a crucial role in enabling early humans to develop tools, cultivate crops, heat homes and eat foods that our stomachs would otherwise be unable to digest. The automobile is another example of a GPT, which increased the productivity of the transport sector by leaps and bounds, allowed people to commute longer distances to work and created new industries, such as for gasoline and motor vehicle parts.

Fossil fuels have also been one of the most transformative GPTs for human progress. Over the last century and a half, they have been the main source of energy for industrial and household purposes. They are used in the production of plastics. They generate electricity, heat buildings and enable greater mobility by allowing people to travel longer distances in shorter timespans. Fossil fuels have been crucial in the growth of modern transportation systems, and they have powered the machines that facilitate the production and processing of food, which has led to increased productivity in agriculture and enabled the world to feed billions of people.

In spite of these enormous benefits, it’s important not to ignore that fossil fuels and other GPTs also have drawbacks. The finite nature of fossil fuels, combined with their negative climate effects, have the U.S. and many other nations searching for alternatives. Likewise, the widespread adoption of the internet and mobile technology has created new markets for e-commerce, digital media and mobile apps, yet social media is bringing new challenges to our culture and our collective psyche. There is an aesthetic loss as well. In the era of email, a well-penned love letter feels like a lost art, and when artificial intelligence can generate thousands of images in seconds, a lot of artwork feels cheap and expendable. That said, we shouldn’t forget that technological innovations bring opportunities for artistic innovation as well. Without the invention of the electric guitar, there would have been no rock and roll.

For better or worse, GPTs unleash change, which creates challenges as some industries are made obsolete and correspondingly suffer job losses. Even as some industries fade, however, others become the sources of new investment and job creation. Businesses reposition themselves to take advantage of these opportunities and earn high returns. Consumers meanwhile respond to technological change by redirecting spending toward new products that improve their welfare. The trajectory of these changes is toward progress. That said, this progress comes with an economic and a human toll, which should not be taken lightly.

The Coming Eon

Today, we may be on the cusp of a new era of technological progress, one in which GPTs will once again play a central role. Fusion energy, artificial intelligence and biotechnology are but a few examples of areas where potential GPTs may be lurking on the horizon. Like fire and fossil fuels, however, these innovations will come with their own challenges and risks.

Fusion energy, for example, has the potential to provide an almost limitless source of clean energy, helping to address climate change and also boosting productivity by reducing the overall cost of energy. However, fusion plants can also leak radiation, and while considered safer than fission plants, they can contribute to proliferation concerns if rogue actors use them to develop fissile materials for the construction of nuclear weapons.

AI meanwhile has the potential to transform industries such as healthcare, finance and transportation by automating tasks, analyzing data and streamlining decision-making processes. But it can also be used in surveillance that erodes our privacy or to generate deep fakes that spread false information, unfairly damage reputations or degrade trust in media.

Biotechnology can possibly revolutionize medicine and agriculture, leading to new treatments for diseases and more efficient methods for producing food. Advancements in biotech are leading to MRNA vaccines, 3D-printed prostheses and safer pesticides that enable greater crop yields. But innovations in this area could also be used in dangerous gain-of-function research that produces new and more dangerous viruses, which could unleash devastating pandemics or be used in biological warfare.

As with all things, GPTs cannot be adopted without tradeoffs. Sometimes the harms are reversible, as with job loss. But other times they are not, as with the roughly 40,000 traffic fatalities that occur each year in the U.S. But this does not mean we are safer if we slow down these advancements. The next GPT might cure cancer, so delaying technological innovation can kill too. Meanwhile, stagnation brings risks of its own. Political divisions and eroding trust in institutions—defining features of modern America—may result from, or at least be exacerbated by, prolonged periods of slow growth.

Embracing Change

Those who resist change may be shouting against the wind, as the future is coming whether we like it or not. Clever entrepreneurs usually find ways to innovate even when legal barriers stand in their way. Regulatory arbitrage, for example, occurs when entrepreneurs seek out countries with legal regimes more favorable to their inventions and innovate there instead of at home.

The best we can do, most likely, is to try and be prepared to the extent we can. That means promoting and encouraging the safe development of GPTs, and anticipating their accompanying risks so we can be ready for them. One thing is for sure, the future is almost never what we expect it to be. The risks of new technologies can be the stuff of nightmares. But if we play our cards right, the benefits can exceed our wildest dreams.