In brief

In brief

  • Where is AI taking us? What will it enable in the future? How will it change industries and business models?
  • It’s important to think about these questions from a perspective informed by historical patterns in the development and adoption of new technologies.
  • Our central example is the rise of refrigeration technology in the 1800s, which we contrast with the rise of AI from the 1950s to present day.
  • By identifying repeating patterns in technological disruption, we give ourselves a better chance to respond to change and avoid classic mistakes.

The rise of many technologies can be traced back through a progression of major breakthroughs and incremental improvements. When we look back, the steps of these progressions often seem logical. Music formats, for example, gradually evolved from analogue to digital formats, moving from vinyl records, to cassettes, to CDs, to Mp3s and now streaming services.

In some cases, it can seem obvious in hindsight how one innovation would naturally lead to the next. But when we try to examine how today’s technologies will progress, the long-term direction is far from obvious. For instance, when Sony released the Walkman™ in 1979, it would have been difficult to imagine Apple’s iPod™ of 2001, let alone the impact of Spotify™ since 2008.

Part of the reason for this is that our historical narratives leave out all the noise, tangents, failures, and second-place finishers. The story of music formats features only the winners, with footnotes, at best, for the likes of Minidiscs, Digital Audio Tapes, Digital Compact Cassettes, Super Audio CD or DVD-Audio.

When we look ahead, we often cannot tell the winners apart from future footnotes—not with any certainty, at least. What we can do, however, is learn from the technology evolutions of the past. Doing so will offer us both a clearer glimpse of the future, as well as a useful perspective on present day decisions.

From artificial ice to artificial intelligence

Leaving music formats for now, let’s turn our attention to refrigeration technology. This began in the 1850s—when the first mechanical ice-making devices were developed—and enters the modern era in the 1950s, when electric fridges became commonplace. The history of refrigeration offers useful parallels with the emergence of AI from its origins after WWII to the present-day proliferation of AI-driven applications and innovations. I’ll draw three key points from these parallels, designed to help you better anticipate the impact and potential of AI in the decade ahead—and beyond.

Expect both revolutions and stagnations

Let’s begin by looking at the pace of development in AI. Often technologies progress incrementally. Think of batteries, vacuum cleaners or running shoes: Over the past 20 years they have all improved, but they still share many core components, features and limitations with their 1999 equivalents.

However, one major breakthrough can see a technology make a giant leap forward in a short space of time. This is what happened with refrigeration in 1928, when the invention of Freon (a refrigerant gas) made refrigerators safe and reliable enough to roll out to mass markets. Today, Freon™, a trade name for chlorofluorocarbons (CFCs), is widely known for causing ozone depletion.

However, one major breakthrough can see a technology make a giant leap forward in a short space of time.

Technologies often go through bursts of innovation—often following a major breakthrough—and then settle again. Touch screen smartphones, too, for example, developed rapidly for a few years after their introduction in 2007 before settling into incremental improvements. Some technologies also go through doldrums where innovation stagnates and investment declines—for example, the “AI winter” of the early '90s and solar power in the 1980s.

AI’s “Freon Moment”

Has AI had a “Freon moment”? Some would argue that the development of deep learning that began to gain pace in 2011 and 2012 is a candidate. This advance in AI-enabled pattern recognition led to Google Photos™ being able to identify your pictures, Siri and Alexa understanding your voice and Netflix™, Spotify™ and Amazon™ knowing what you may like.

The uncomfortable lesson is to always be ready for rapid change from AI innovation—and to constantly try to anticipate its direction. But equally, we need to be aware that these surges can occur after long periods of slow progress. We should resist the urge to think that stagnation, steady improvements, or rapid change will continue at the same rate.

Look for creative combinations and new enablers

Technologies often advance via the development of complementary technologies and combinations of innovations from diverse industries, functions and segments.

The transistor, for example, was developed by the telecommunications industry in 1948. It was certainly a "Freon moment" for that industry, but it also went on to revolutionize electronics of all kinds and paved the way for all subsequent computer technology.

The rapid progress of AI since 2011 has only been possible because of advances in computing power, cloud computing, enterprise IT, and APIs—none of which have been developed purely to drive AI forward. It was no different for refrigeration, where electrification, mass production, and distribution were all needed to support the explosion of household electric fridges in the 1950s.

When we think about where AI is going, we should also think about where dozens of other technologies are going—and how these could support or combine with AI to drive transformation. For example, quantum computing is seen as a potential game changer for AI by crunching data at unprecedented levels of efficiency. Technological advancements could also come from developments not strictly related to AI too, such as innovations in data privacy or breakthroughs in how we understand the biological computation of neurons and brains.

The rapid progress of AI since 2011 has only been possible because of advances in computing power, cloud computing, enterprise IT, and APIs—none of which have been developed purely to drive AI forward.

When change is inevitable, do something new

Will your company or industry spend the next ten years battling to survive disruption—or lead the direction of disruption?

Many people may look back at companies like Kodak™ and Blockbuster™ and conclude that they weren’t able to anticipate the world evolving around them. But this is not accurate. These companies did try to change. They were certainly more reactive than innovative, but their failures were more because they could not transform than because they would not.

When the ice empires melted

Companies have been struggling with transformation for centuries. The ice trade, which began in 1806, was pioneered by Frederic Tudor of New England. He packed ships full of ice from frozen lakes and transported it from the Northeast United States to destinations around the world (using only sawdust for insulation).

Tudor became a millionaire from his ice trading empire, and his model was the blueprint for the birth of a major global industry. The ice trade had its heyday in the decades after 1880, with the volume of ice traded tripling between 1880 and 1907.

But disruption was looming. In 1844, American inventor, John Gorrie, developed the first mechanical ice-making machine. Throughout the 1850s, inventors around the world—from James Harrison in Australia to Ferdinand Carré in France—developed improvements and alternative designs.

This was the dawn of what was then called, “artificial ice” (another “AI”, coincidentally). And from 1870 to 1914, while the “natural ice” industry was thriving, artificial ice machines secured a foothold in two major industry groups: breweries and meat-packing plants.

Avoid merely polishing yesterday’s business

People in the natural ice industry could see there was a genuine threat to their businesses. Some reacted by investing in new tools that would speed up the rate of ice harvesting, but these improvements were little help against rapidly advancing refrigeration technology. Like many companies facing disruption, these organizations spent time focusing on how to run their current business more efficiently, rather than making the drastic changes needed to survive. In the end, some companies survived by entering new industries (including artificial ice production), but the global natural ice trade industry collapsed and vanished soon after WWI.

You can make a strong case that we live in similar times today—that the way we approached decision- and prediction-making prior to AI is analogous to how the natural ice industry kept things cold prior to refrigeration. Much of AI is about embedding automated decisions and predictions into the fabric of our lives. These have always relied on human minds, just as cooling and freezing always relied on natural ice and weather before refrigeration.

The bottom line is that many of us need to spend more time working out how to make a transformative transition to an AI-dominated world, instead of making incremental improvements to the way things are now.

Today, AI is driving a revolution in how we predict and automate outcomes and decisions. No matter what industry you are in, the AI-driven mechanization of decisions and predictions—much like the mechanization of cooling and freezing—will drive extensive disruption in the decade ahead.

The bottom line is that many of us need to spend more time working out how to make a transformative transition to an AI-dominated world, instead of making incremental improvements to the way things are now.

Building a business for tomorrow

I will leave you to imagine where is AI taking us and what it will enable in your industry, but I urge you to do this with a perspective informed by the long history of disruptive technologies.

Just from the handful of examples in this article, we can learn to expect dynamic rates of development in AI—anticipating the next “Freon moment” while accepting it may come in the next month, the next decade, or not at all. We learn to take a broader look at human progress when thinking about AI, understanding how combinations of enabling factors and new technologies will drive most breakthroughs. Electrification made home fridges and freezers possible, high-speed internet enabled streaming music services—so let’s ask ourselves: What will be the next major enabler of progress for AI?

Perhaps most importantly, we learn not to be drawn into the (often more comfortable) task of making yesterday’s business more efficient at the expense of building a business for tomorrow.

When we look back fifty years from now, the change that AI drives will likely look as obvious as the change that refrigeration drove in the 1920s. But by drawing on insights from the history of innovation and technology, we stand a greater chance of driving and pre-empting disruption, instead of passively awaiting change.

Subscription Center
Stay in the know with our newsletter Stay in the know with our newsletter