Will AI Solve the Business Productivity Paradox - Yes - BUT….

Introduction

We are in the eye of the hurricane as we see huge transformational technologies emerge in the world of Artificial Intelligence and particularly within that Generative AI, based on Large Large Models (LLMs) and General Pre-Trained Transformers (GPT), which provide seemingly unlimited access to insights from across the public internet. In the 1970s to 1980s a productivity paradox was defined as a perceived "discrepancy between measures of investment in information technology and measures of output at the national level." The concept is attributed to Robert Solow, in reference to his 1987 quip, "You can see the computer age everywhere but in the productivity statistics."

While the computing capacity of the U.S. increased a hundredfold in the 1970s and 1980s, labour productivity growth slowed from over 3% in the 1960s to roughly 1% in the 1980s. Productivity growth spring into life for the next 20 years, but has been largely moribund again since the Global Financial Crisis of 2007-8, despite continued significant investment in Computing Power, Digital and now Artificial Intelligence.

One critical source of Productivity Growth is Energy. The widespread use of Coal and then Oil to drive steam turbines and subsequently generate electricity was the fundamental enabler of productivity growth in the first Industrial Age. Whilst in the 1950s Nuclear Power promised a further exciting step change in growth capacity, this has not delivered returns not least because of health and safety issues. Nuclear Fusion providing ‘limitless’ energy and more recently the demands of NetZero migrating carbon fuels to more sustainable sources has again restricted the growth in energy capacity in Western Economies. Furthermore, the exponential growth in Computing Capacity is driving ever more demand for sources of power generation.

The question therefore remains, can we produce a step change in productivity and economic growth? My simple answer is Yes, BUT!

No…Yes

With all of the excitement around AI, MIT Professor Daron Acemoglu, recently released “The Simple Macroeconomics of AI”, with the aim to review the large macroeconomic implications of new advances in AI. Acemoglu using existing estimates on exposure to AI and productivity improvements at the task level, these macroeconomic effects appear nontrivial but modest—no more than a 0.66% increase in total factor productivity (TFP) over 10 years. starts from a task-based model of AI’s effects, working through automation and task complementarities, and further revises this down to 0.53%.

Meanwhile, Professor Erik Brynjolfsson who is a powerful advocate of the productivity potential that can be achieved through AI. He has written a number of seminal books including The Second Machine Age, and Machine, Platform, Crowd with Andrew McAfee. Brynjolfsson documented a correlation between IT investment and productivity, going back to his earliest research in 1993 at the behest of Robert Solow. His work provides evidence that the use of Information Technology is most likely to increase productivity when it is combined with complementary business processes and human capital. A subsequent article coined the term the Productivity J-Curve to describe how these intangible investments might initially lead to stagnant or even lower productivity followed by a take-off.

Brynjolffson envisions double-digit gains in economic productivity, and commenting that he foresees the potential for “massive economic disruption” leading to the creation of new occupations and new companies in the coming years. In an interview for the Financial Times he said “I’m optimistic the technologies will affect a large number of tasks. A big percentage of the work that is done in a modern economy is amenable to being augmented by LLMs and generative AI. The effects on those tasks have been significant — double-digit productivity gains within just a few months in some of the cases I’ve studied. Multiply the large percentage of affected tasks by sizeable productivity gains for each one and you get a big total economic impact. I’m betting that productivity growth is maybe significantly higher in the 2020s than the Congressional Budget Office is projecting. They projected 1.4 per cent average per year. I think it could be twice that — closer to 3 per cent — maybe more.”

But… Yes – The Intelligent Enterprise

How do we square these different research opinions.

Looking at the paper from Prof Acemoglu, if we treat AI like other ‘automation’ initiatives and focus on micro-tasks then the costs could well outweigh the benefits! Generative AI is based on very expensive compute power and the danger is we throw it any potential task in the organisation. In my opinion, this will be the source of risk and failure for any companies working with the Big Strategy and Consulting firms…

Reading the work of Brynjolfsson, we need to rethink how work is done. In his work on the Productivity JCurve, he describes how General purpose technologies (GPTs) such as AI enable and require significant complementary investments, including co-invention of new processes, products, business models and human capital. These complementary investments are often intangible and poorly measured in the national accounts, even when they create valuable assets for the firm.

I have written many times that the majority of current company operating models are based on the 19th Century manufacturing process-centric construct. This model was automated/industrialised in the 1990s and 2000s with ERP and CRM systems, following a People/Process/Technology framework, which was largely Process drives Technology and People are the business case. Whilst this model did deliver the growth in Productivity that was long overdue it ran out of road 15 years ago.

We now need to go much further, revisiting the ‘Reengineering the Corporation’ of Michael Hammer and instead of automation processes, ‘obliterating tasks and processes’, consider the ‘Good to Great’ Capabilty Flywheel and start to use Data and AI to dynamically reconfigure tasks and eliminate processes. Eliminating activities and doing new things quite differently, in order to spin your ‘Capability FlyWheel’ faster, is the art of the #intelligententerprise, a business that is fully data enabled and AI Powered.

That is the essence of the graphic at the top of this article. If we aim to implement AI in the same old/same old Operating Model construct, you will get the <1% improvements promised by Acemoglu. If you aim to rethink how your business works to embrace AI and become an Intelligent Enterprise, the possibilities are almost endless.

How you get there is the topic of the Seven Habits of the Highly Effective Data Enabled, AI Powered Business, more of which is available on this Blog.

1. Acemoglu, Daron, The Simple Macroeconomics of AI, MIT, May 2024

2. Brynjolfsson, Erik (1993). "The productivity paradox of information technology". Communications of the ACM. 36 (12): 66–77. doi:10.1145/163298.163309. ISSN 0001-0782. S2CID 15074120.

3. "The Second Machine Age". May 20, 2023.

4. "The long wait for a productivity resurgence". www.ft.com. Retrieved March 8,2024.