In this paper we evaluate estimations of Generative AI's economic impact with those of other technologies from both modern and historical times in order to place forecasts of Generative AI's influence into scale. Since these are the two most common measures given by the economic potential of generative ai reports and offer simple to compare estimates, we evaluate impact along two dimensions: GDP and productivity.
Background
One of the main engines of economic development is technological improvement. Many times, these developments follow waves—that of the Industrial Revolution and the Information Revolution. Other times, single, significant technologies fuel the economic potential of generative ai booms and productivity increases. Often cited examples are the current computer, the Internet, electricity, the steam engine, and—backward even more—the printing press. Generative artificial intelligence is the newest supposed to be breakthrough technology.
With the publication of GPT-3 in 2020 (and shortly-to-follow ChatGPT in November 2022), which sparked the present AI-arms race among technology companies and pushed discussion and use of AI tools to the wider public, the possibility for such transforming impact started. Writings and arguments on the financial influence of this technology flourish alongside GenAI products derived from these ideas. With the likelihood of today's tasks being carried out mostly by GenAI agents in the near future, expert claims of Generative AI's consequences cover the range from a notable increase to the GDP of developed economies to a complete and utter restructuring of the global labor market.
GenAI estimates reveal a lot of uncertainty and great possibility for impact.
Though Generative AI's possible implications on the economy have generated a lot of debate, there have been shockingly few particular, quantitative projections published on the scale of its economic influence. Among the limited projections made, the estimations also often show substantial variation. With a median projection of under $1 trillion yearly over the next ten years, Accenture's 2024 research and Goldman Sachs' 2023 report have presented the lowest estimates for impact on global GDP. Forecasting an increase in GDP of anywhere from $2.6 to $4.4 trillion annually - a 2x to 4x bigger impact than the low-end predictions, at the top end is an often referenced McKinsey 2023 report.
The GDP increase of generative artificial intelligence could match those of the biggest nations on Earth.
McKinsey's 2023 report, with its maximum end value, would show a worldwide GDP rise more than that of the sixth-largest economy in the world, the UK economy right now. A significant growth in the size of the Netherlands' or Saudia Arabia's economy, annually would be a result of a $1 trillion annual gain in added value (not simply income) to global GDP even on the low end of the predictions.
Generative artificial intelligence also greatly increases productivity.
Examining the expected consequences on production, we find similarly significant changes that would cause a global revolution in economies all around. In the years before COVID, global productivity increase among OECD countries varied from 0.5 to 1.6%. With the mean projections reflecting a doubling of yearly productivity gain owing to Generative AI, Goldman Sachs' and McKinsey's predictions for GenAI productivity effect show potential for ranging from 0.1% to 2.9% annual growth.
Generative artificial intelligence is, all things considered, expected to have a significant financial impact. Comparatively to other technologies, including those also now projected and others that have been measured, GenAI forecasts assist one to put the impact of this technology into perspective.
2010s AI economic effect estimates surpass Generative AI's
Artificial intelligence forecasts covering more conventional application cases of AI existed prior to the publication of GPT-3 and ChatGPT. By contrasting these "pre-GPT" estimations of AI's influence with the new GenAI projections, one may get the most clear image of the potential size of impact the new Generative AI technology could have over the current technological scene.
Fascinatingly, artificial intelligence forecasts for the 2010s are far greater than present Generative AI effect estimates. Regarding GDP, GenAI matches the projection of Analysis Group and has the same scale as the other estimates. GenAI appears indifferent from 0% when placed on the same scale as the AI forecasts, far from the productivity effect value. This suggests to us that rather than the channel of increasing productivity, the projections for the value of GenAI are based on the cost savings possible of the technology.
Other technology projections line more closely with GenAI's projections.
The Internet of Things (IoT), cloud technology, and driverless cars are just a few of the many other practical benchmark technologies with the economic potential of generative ai. Though many of these technologies are still in their early years of use, like Generative AI they are expected to have significant effects. With the mean of the Gen AI estimates falling precisely between advanced robotics and next-generation genomics, GenAI's GDP impact seems to be smack in the middle of a spectrum of current-day technological innovations. Although many of these projections have a broad range, it seems that forecasters who are guilty of being overly optimistic about GenAI's future are equally guilty of such optimism with other technologies.
Generative artificial intelligence also lies midway between the effects of previous technology.
Examining the actual, quantifiable impact of earlier technologies provides another way we might scale Generative AI given our comparisons of estimates of its economic influence to other current technologies. Although it is challenging to determine the exact influence of a given technology across time, some studies have extracted individual effects by means of statistical approaches. Among these are such classic technologies like manufacturing robotics and the steam engine. Combining GenAI projections with the actual worth of earlier innovations provides another yardstick for how closely such forecasts might fit reality.
Once more, the GenAI forecasts show good fit with the observed effects of early 21st-century technology as opposed to those of 20th century ones. On the high end of the estimates, GenAI might offer more than double the productivity increase of the Information and Communication Technologies (ICT) and Information Technology (IT) developments of the early 2000s. On the other hand, the bottom end of the productivity estimates puts GenAI's influence just below these technologies.
In essence
The comparison between Generative AI and other technology forecasts reveals that, if one thing stands out, it is still a lot of ambiguity regarding GenAI's influence. Forecasts span the situation in which Gen AI may revolutionize the global economy contributing annual value equivalent to the sixth-largest GDP worldwide, all the way to the scenario in which it is hardly a blip on the radar compared to past AI technologies.
Luna Ella
In this paper we evaluate estimations of Generative AI's economic impact with those of other technologies from both modern and historical times in order to place forecasts of Generative AI's influence into scale. Since these are the two most common measures given by the economic potential of generative ai reports and offer simple to compare estimates, we evaluate impact along two dimensions: GDP and productivity.
Background
One of the main engines of economic development is technological improvement. Many times, these developments follow waves—that of the Industrial Revolution and the Information Revolution. Other times, single, significant technologies fuel the economic potential of generative ai booms and productivity increases. Often cited examples are the current computer, the Internet, electricity, the steam engine, and—backward even more—the printing press. Generative artificial intelligence is the newest supposed to be breakthrough technology.
With the publication of GPT-3 in 2020 (and shortly-to-follow ChatGPT in November 2022), which sparked the present AI-arms race among technology companies and pushed discussion and use of AI tools to the wider public, the possibility for such transforming impact started. Writings and arguments on the financial influence of this technology flourish alongside GenAI products derived from these ideas. With the likelihood of today's tasks being carried out mostly by GenAI agents in the near future, expert claims of Generative AI's consequences cover the range from a notable increase to the GDP of developed economies to a complete and utter restructuring of the global labor market.
GenAI estimates reveal a lot of uncertainty and great possibility for impact.
Though Generative AI's possible implications on the economy have generated a lot of debate, there have been shockingly few particular, quantitative projections published on the scale of its economic influence. Among the limited projections made, the estimations also often show substantial variation. With a median projection of under $1 trillion yearly over the next ten years, Accenture's 2024 research and Goldman Sachs' 2023 report have presented the lowest estimates for impact on global GDP. Forecasting an increase in GDP of anywhere from $2.6 to $4.4 trillion annually - a 2x to 4x bigger impact than the low-end predictions, at the top end is an often referenced McKinsey 2023 report.
The GDP increase of generative artificial intelligence could match those of the biggest nations on Earth.
McKinsey's 2023 report, with its maximum end value, would show a worldwide GDP rise more than that of the sixth-largest economy in the world, the UK economy right now. A significant growth in the size of the Netherlands' or Saudia Arabia's economy, annually would be a result of a $1 trillion annual gain in added value (not simply income) to global GDP even on the low end of the predictions.
Generative artificial intelligence also greatly increases productivity.
Examining the expected consequences on production, we find similarly significant changes that would cause a global revolution in economies all around. In the years before COVID, global productivity increase among OECD countries varied from 0.5 to 1.6%. With the mean projections reflecting a doubling of yearly productivity gain owing to Generative AI, Goldman Sachs' and McKinsey's predictions for GenAI productivity effect show potential for ranging from 0.1% to 2.9% annual growth.
Generative artificial intelligence is, all things considered, expected to have a significant financial impact. Comparatively to other technologies, including those also now projected and others that have been measured, GenAI forecasts assist one to put the impact of this technology into perspective.
2010s AI economic effect estimates surpass Generative AI's
Artificial intelligence forecasts covering more conventional application cases of AI existed prior to the publication of GPT-3 and ChatGPT. By contrasting these "pre-GPT" estimations of AI's influence with the new GenAI projections, one may get the most clear image of the potential size of impact the new Generative AI technology could have over the current technological scene.
Fascinatingly, artificial intelligence forecasts for the 2010s are far greater than present Generative AI effect estimates. Regarding GDP, GenAI matches the projection of Analysis Group and has the same scale as the other estimates. GenAI appears indifferent from 0% when placed on the same scale as the AI forecasts, far from the productivity effect value. This suggests to us that rather than the channel of increasing productivity, the projections for the value of GenAI are based on the cost savings possible of the technology.
Other technology projections line more closely with GenAI's projections.
The Internet of Things (IoT), cloud technology, and driverless cars are just a few of the many other practical benchmark technologies with the economic potential of generative ai. Though many of these technologies are still in their early years of use, like Generative AI they are expected to have significant effects. With the mean of the Gen AI estimates falling precisely between advanced robotics and next-generation genomics, GenAI's GDP impact seems to be smack in the middle of a spectrum of current-day technological innovations. Although many of these projections have a broad range, it seems that forecasters who are guilty of being overly optimistic about GenAI's future are equally guilty of such optimism with other technologies.
Generative artificial intelligence also lies midway between the effects of previous technology.
Examining the actual, quantifiable impact of earlier technologies provides another way we might scale Generative AI given our comparisons of estimates of its economic influence to other current technologies. Although it is challenging to determine the exact influence of a given technology across time, some studies have extracted individual effects by means of statistical approaches. Among these are such classic technologies like manufacturing robotics and the steam engine. Combining GenAI projections with the actual worth of earlier innovations provides another yardstick for how closely such forecasts might fit reality.
Once more, the GenAI forecasts show good fit with the observed effects of early 21st-century technology as opposed to those of 20th century ones. On the high end of the estimates, GenAI might offer more than double the productivity increase of the Information and Communication Technologies (ICT) and Information Technology (IT) developments of the early 2000s. On the other hand, the bottom end of the productivity estimates puts GenAI's influence just below these technologies.
In essence
The comparison between Generative AI and other technology forecasts reveals that, if one thing stands out, it is still a lot of ambiguity regarding GenAI's influence. Forecasts span the situation in which Gen AI may revolutionize the global economy contributing annual value equivalent to the sixth-largest GDP worldwide, all the way to the scenario in which it is hardly a blip on the radar compared to past AI technologies.