The Evolution of the Business Buzzword, ‘AI’

by Michael Watson

“AI,” or Artificial Intelligence, is a very popular tech buzzword that seems to take on a new meaning each and every time you hear it. This causes great confusion for many business leaders struggling to put it into perspective for their own specific needs.

The truth is: AI has definitely evolved into an umbrella term used to encompass a wide variety of algorithms and approaches. This comes as no surprise, however. In a rapidly moving field with so many tools and algorithms that seem to just ‘fit’ within the numerous definitions of “intelligence,” it seems logical to include them under the same umbrella. This doesn’t help alleviate the confusion, however. In order to help you clarify the term for your own use, let’s go back in time and see how we got where we are today.

The term “analytics” became popular due to the 2006/2007 Harvard Business Review article and book by Tom Davenport called “Competing on Analytics.” The content made companies realize they should be doing more to ensure they were making data-driven decisions across their entire enterprise. Davenport made the case that if companies weren’t using data to continually make better decisions, they risked losing out to competitors who did. Business leaders took this message to heart and the term “analytics” took off.

For a while, the term “analytics” was co-opted to refer mostly to data reporting (also called Business Intelligence), which greatly diluted its potential value. As important as data reporting is, it can only take you so far — there is much more to using data than just dashboards! As a reaction to this, serious thinkers presented us with the terms:

“Descriptive Analytics” (which includes reporting)

“Predictive Analytics” (forecasting the future or classifying data)

“Prescriptive Analytics” (using data to suggest actions)

With these terms, we then had a more sophisticated way to think about obtaining value from data, well beyond reporting and business intelligence systems.

Around 2014, the term “machine learning” became a more recognized term in business settings as companies began to realize the value these new, readily available, algorithms could provide. The business community knew that machine learning algorithms could lead to better predictions, or even to predictions in completely new areas. This created a lot of new business value. And yet, the term “machine learning” always referred to the algorithms; it never became an umbrella term.

However, back in 2012 and unbeknownst to the business world, the term AI began gaining steam in the research community. That year, there was an important breakthrough in algorithms for image recognition. Part of this was because of improvements in hardware — specifically the use of GPUs for massive parallel computing, but another big driver was advances in deep neural networks (really just advanced machine learning algorithms). In 2012, for the first time, it became clear that these deep neural networks could recognize images better than people could. Researchers and application developers doubled-down on this approach and invested heavily.

Historically, people give credit to the 1956 Dartmouth Conference for the original rise of the term AI. The term has come in and out of favor in the intervening decades with a variety of different approaches claiming to be on the verge of AI. In 2012, researchers had a good reason to revive the term AI: the underlying algorithms were neural networks. Neural network algorithms are loosely based on our current understanding of how the brain operates (through a network of neurons). If these algorithms truly do work like the human brain, then the term “AI” is a natural (and fun) way to describe the algorithms.

From 2012 onward the research efforts paid off and the advances were impressive: better image and video recognition, realistic autonomous vehicles, great language translation and understanding plus well-publicized victories in board and video games. By 2017, it was clear these advances could dramatically change many industries and alter how a business was run. The term really caught on, and it captured our imaginations.

At the same time, more algorithms were placed under the umbrella term “AI” — even if the algorithms didn’t use a neural network. Of course, as the term “AI” became more of an umbrella term in the general business community, the research community found a need to differentiate it from the ongoing goal of building machines that “think” in ways more human-like. Today, the research community sometimes uses the term “artificial general intelligence” (AGI) to describe systems that learn and react just like a human. Think of this as the self-sufficient robots of science fiction. In our view, AGI is still in the very early stages.

This leaves the term “AI” (or “narrow AI” in some circles) for the technology that exists today. For example, when an AI algorithm identifies an image as a cat, we don’t expect general intelligence from that algorithm to decide whether to feed or take the cat to the vet.

Overall, this wide reaching definition is a good thing. Used in this way, the term AI better reflects how you should be thinking about the various algorithms and new advances that are now being implemented by the world’s leading businesses.

In our view, the use of “AI” as an umbrella term keeps things simple by embracing many types of algorithms rather than debating where an algorithm may fit or creating new terms like “advanced analytics” to differentiate it from reporting.

We also think “AI” is a great umbrella term because it represents a greater call for action than the term “analytics” did. “Analytics” perpetuated a misinformed view that reports, dashboards, and insights might be enough while “AI” clearly implies that you have to think more carefully about how to change your business, your workflows and the jobs people do.

In other words, if you are not reinventing your business with AI, your competitors will be.

In the end, the most important realization to be aware of is that the definition of AI that suits your business may be entirely different for others. For example, computer vision may be an essential AI advancement at Google while Amazon’s key AI application could be demand forecasting. Your business might be transformed by a different AI algorithm altogether. The bottom line is that AI is behind many of the most influential transformations in business but exactly how it is applied in each instance is unique.

The power behind AI is what you make of it.

If you liked this blog post, check out more of our work or follow us on Social Media.