Robert B. Cohen, Ph.D., Senior Fellow, Economic Strategy Institute, May 11, 2018
(click here to download the full report in PDF format)
1. An Initial Estimate of the Impact of AI and ML on the U.S. economy
The impact of AI and ML on the economy is becoming clear through corporate presentations about how they are introducing these technologies. First, AI and ML will alter how firms operate and manage automation within their environments, whether that means manufacturing or services. Second, AI and ML will support a substantial evolution in software, infrastructure and computing for businesses, based largely on innovations in computing such as the move to Cloud Native (where no operating system is required on computers used to work in clouds). Part of this evolution includes virtualized infrastructure as well as the emergence of orchestration tools such as Kubernetes that accelerate the pace of application development. Third, with these enhancements, businesses can make more effective use of AI and ML applications to manage machines on the factory floor as well as support data analysis. They can also better manage human resources, supply chains, procurement, and other business functions.
We have estimated that current spending on AI, ML and related data analysis creates about $400 billion to $940 billion in output, thus raising GDP by about 1% to 5%. We think that recent improvements in business operations that result from improved cloud computing, more rapid development of software and the Internet of Things, are being driven by hiring and spending levels that we believe we can forecast. For the period from 2015 to 2018, we estimate that large tech “superstar” firms, such as Facebook, Amazon, Netflix, Google, Apple, NVIDIA, Walmart, Capital One and others have hired 40,000 to 120,000 people in data analysis roles where AI and ML are involved. Using a similar rough estimate, we think that other traditional firms have hired about 50,000 to 200,000 people working in areas related to AI and ML over the same period. In addition, we estimate that about 50,000 to 200,000 people are working in small, startup firms that are using AI and ML in products and services they are developing.1 In total, we estimate that from 2015 to 2018, firms in these categories have added 190,000 to 520,000 jobs to the economy. We also estimate that the large tech firms probably spent about $20 billion to $120 billion to support the move to AI, ML and data analytics, that the large traditional firms spent $80 billion to $240 billion, and that startups, numbering 10,000 to 20,000 firms, have spent $$20 million to $100 billion each. This would result in spending of $200 billion to $2 trillion. In sum, this would indicate there has been $200 billion to $2.36 trillion in spending over the 2015 to 2018 period. This would average out to about $70 billion to $780 billion a year. If we expect this spending to contribute to GDP in a ratio of $2 for every dollar of spending, this would mean that in one year, GDP would increase by $200 billion to $1.5 trillion a year. This would translate into an estimated 1% to 7% increase GDP in 2018. (which we assume is about $20 trillion). This is probably a high estimate, but since we are including not only AI and ML, but also data analysis and software development, the figures would be higher than spending on AI and ML alone.
We think that these estimates are for the very first stage of a substantial change in software and infrastructure. We expect that the contribution of AI and ML to the economy will grow rapidly over the next few years
Click here to read the full report in PDF format.