The Economic Potential of Gen AI: Balancing Expectations and Realities

The debate around Generative AI (Gen AI) is polarizing, with viewpoints ranging from overly optimistic to deeply skeptical. As a Digital Transformation practitioner who has implemented numerous Gen AI use cases at Automation Anywhere, I aim to provide a grounded perspective.

Goldman Sachs, in Global Macro Research publication, (Jun 24) featured interview with Daron Acemoglu, MIT Institute Professor who has published numerous books including Why Nations Fail: The Origins of Power, Prosperity, and Poverty and his latest, Power and Progress: Our Thousand-Year Struggle Over Technology and Prosperity. This also included interview from Joseph Briggs, Goldman Sachs Economist and Goldman Sachs analysts.

Daron Acemoglu is skeptical about the benefits touted by many for Gen AI. He argues that Gen AI will provide a modest 0.9% cumulative GDP growth over 10 years, roughly 0.1% growth per year. His assumption is based on the fact that only 5% of corporate tasks can be automated, resulting in 0.6% productivity increase.

This contrasts with Goldman Sachs economist Joseph Briggs, who predicts overall 25% of tasks could be automated leading to 6.1% cumulative GDP over 10 years, roughly ~1% growth per year.

McKinsey, in its article Economic potential of generative AI: The next productive frontier, has identified 60 use cases, potentially contributing $2.2T to $4T annually. With additional productivity use cases, this could rise to $6-$7T per year.

Given the uncertainty, the spectrum ranges anywhere from $0.1T to $7.9T a year. How do we size this opportunity?

Comparison to Other Revolutionary Technologies

Let's look at another exponential technology and how it impacted the world growth. Can we draw some conclusions from this? Even though these technologies are not parallel, hopefully it can help us to understand the projections and estimations.

The internet revolutionized productivity growth in the early 21st century. From 1995 to 2022, the top 10 economies' GDP grew from $22T to $67.3T. Limiting the analysis to the top 10 economies (accounting for 68% of Global GDP) and excluding China (whose growth was powered by globalization), the top 9 economies grew from $21.4T to $50T, a 3.1% CAGR. The internet acted as a catalyst for this growth.

If Gen AI drives similar growth, it could contribute ~$2T annually to GDP of Top 10 economies. However, this growth won't solely be due to Gen AI but also other technological advancements that could contribute productivity growth including AI/ML.

Gen AI Adoption

In my article last week, "Gen AI Adoption Spikes, but Not Everyone Is Reaping the Benefits: What Is Wrong?", I analyzed that Gen AI technology is currently being used mainly for foundational use cases like summarization and translation, resulting in low value. However, from my experience implementing several transformative use cases at Automation Anywhere, I can confidently say that Gen AI, when coupled with automation, can deliver significant results. As some of the readers commented, many companies are still trying to evaluate the technology and understand its potential.

Gen AI is revolutionary. However, an annual contribution of $7T (~10% GDP growth per year) seems overly optimistic compared to similar technologies like the Internet. Conversely, an annual contribution of $0.1T seems overly pessimistic.

Based on two methods, my estimation is approximately $1T - $1.5T annual GDP contribution, which is slightly higher than Goldman Sachs' Joseph Briggs's estimation of $0.5-$1 trillion.

Do you agree with this perspective on Gen AI's economic impact?