The AI hype: can consulting firms trust it?

There’s no denying that AI has taken the world—and, more specifically, the consulting industry—by storm, but what happens after the hype? 

We brought together experts from leading AI-focused management consulting firm, KUNGFU.AI, to find out their predictions for the future of artificial intelligence and how it will potentially shape the world of consulting. 

Christian Barnard (COO) and Ron Green (CTO) are joined by Prof. Joe O’Mahoney, Professor of Consulting at Cardiff University, for some wider industry insight and a closer look at the past and present trends of consulting… to see into the future.

‍

A brief history of AI 

At this point, AI is almost impossible to avoid—you’d be hard-pressed to find a consulting firm that hasn’t got at least some familiarity with popular generative tools like ChatGPT. However, as Ron points out, few people realize just how far back the systems behind these tools go. All the way back to the 1940s, in fact.

“It stemmed from an attempt to mimic certain aspects of the brain,” he explains. “It’s seen trillions of dollars of research and experimentation over the years, but there’s been a lot of overpromise and underdelivering. Until about eleven years ago.”

That’s because 2012 was the year that AI experts unlocked the capabilities of machine learning. Machine learning is the branch of artificial intelligence that all the modern systems we use today were derived from. It focuses on teaching computers to learn and make decisions through the analysis of data, all without being programmed.

Ron explains further: “Within the subset of machine learning, there’s also the subset of deep learning. This uses artificial neural networks to build systems that are capable of performing the impressive feats we know today, and very quickly.”

In simple terms, deep learning machines take data in and spit outputs out. Now, that output initially can be completely random or totally incorrect—but then you correct the model. And then you repeat the process. And then you do it many, many more times until it not only learns to answer the questions that it’s asked—it also learns to generalize.

That takes the model to the point at which it makes predictions based on data it’s never even seen before, which is the level of sophistication that we see in today’s tools, like ChatGPT. But while ChatGPT is undeniably amazing technology, it’s only one of countless AI tools available on the market. 

“ChatGPT is just the tip of the iceberg,” says Ron. “Some systems can predict sales trends. Some can make product recommendations. Some diagnostic AI tools are even trying to understand health data—there’s just so much happening in the world.”

‍

Are we in an AI hype cycle?

The velocity of AI algorithms have increased tremendously in recent years thanks to increased computing power and the availability of big data. As a result, research and R&D into AI has also increased, which has led to investment and facilitated growth. 

Yet Ron, who has been building AI-based systems since 1992, believes we’re still only on day one when it comes to realizing the full potential of this technology.

“This moment in time is analogous to some of the most important discoveries in human history, like fire or the wheel,” he insists. “The things we always felt we were on the cusp of doing, we can finally do.”

So, what makes AI so impactful? The answer lies in its exponentiality. 

To take Ron’s comparison to inventions like fire or the wheel—as critical as these technologies were, they weren’t able to build on top of themselves like AI can. The AI systems we have today are the basis for building even more powerful AI systems, and could unlock entire new vistas we’ve still yet to consider.

“It’s very difficult to predict where we’ll be in even ten years due to the pace of the acceleration,” Ron says. 

Professor Joe O’Mahoney agrees that AI’s capabilities are continuously improving at an exponential rate: “Six months ago, it could beat a doctor at spotting cancer from a scan. Now, it scores higher in empathy than the average doctor.”

With language learning models that can translate spoken text to any number of different languages at a time, the ability to generate images instantly based on a line of text, and everyday advancements like predictive text, there’s no signs of AI and its advancement slowing down just yet.

Watch the full AI discussion here.