Brian Pentland is the Main Street Capital Corporation Intellectual Capital Endowed Professor in the Department of Accounting and Information Systems. Additionally, he has served as a faculty member in the Broad Executive MBA, enhancing the curriculum through organizational learning. In this blog post, Pentland shares his insights on leadership and technology trends.
In the world of information systems, there is always a lot of hype on the horizon. Back in the 1990s, when I was first teaching information systems, AI was all about “rule-based systems” and “knowledge engineering.” There was a ton of hype, but the results were pretty disappointing.
More recently (around 2016, as I recall), I had students telling me that autonomous vehicles would be on the road, in general use, right around the corner. We are still waiting, and the corner seems to be getting farther away, not closer. And nowadays, we are hearing about quantum computing and liquid neural networks and the metaverse. How can we sort through the hype and focus on technology that matters?
That’s a reasonable question, but it implies a kind of passive mindset — as though we are sorting through technologies as given. We need to be more proactive. Organizations are in a position to develop and adapt new technologies; they can invent new processes and business models. So, maybe a better question is this: Which technologies will co-evolve with new business models and business practices in a competitive, sustainable way? It’s still not an easy question, but at least we are not passive bystanders.
For example, the most important information technology on the horizon these days is AI in all its many forms. AI and machine learning are definitely not just hype. Systems based on neural networks have been outperforming humans on specific tasks for several few years, and they are getting better fast. There are substantial new developments all the time. Machine learning and AI can be used anywhere there is data — and data is everywhere. Because the potential applications are so pervasive, people make an analogy between AI and the internal combustion engine or electricity. Transformative! Yet, at the same time, outside of some niches, machine learning and AI have had surprisingly little impact on the typical business.
slideshow has a previous and next buttons
So, we come to a puzzle: Why isn’t this transformative technology more transformative?
I would argue that the difference has to do with organizational and managerial factors. For a whole host of reasons, organizations find it difficult to implement new routines and new business models. Managers need to push forward, but technological change is an industry-wide phenomenon. The organizational ecology needs to evolve to support the new value chains. There are strategic issues and human resource issues, as well as technical issues. I suspect that in most cases, the organizational and managerial issues are more complex and require more creativity and innovation than the purely technical issues.
So, the answer to this puzzle is simple: Technology trends are not just technical. It’s organizations (and their managers, their suppliers and their customers) that make the trends.
Let’s take this whole idea a little bit further. Machine learning depends on data. Some say data is the new oil. What does it take to make oil valuable? Crude oil by itself is just a sticky, horrible mess. But if you can locate it, drill it, pump it, transport it and store it, then you have something that really is useful. You can even make it into plastics. There are a lot of applications for oil.
So, how do we make data valuable? Like oil, you need to be able to locate it, select it, refine it and put it to use. You need a pipeline, so to speak. The traditional mindset back in the 1990s was that data is just for decision making. Better data, better decisions. In a static world, maybe that’s good enough, but the current mindset is that data is for learning. Data is for adaptation. To make that happen, to extract value from the data, we need learning systems. Like machine learning systems, for example.
While that sounds good, these neural networks don’t train themselves. To be useful, machine learning requires individual and organizational learning. It requires people to figure out how to adapt to it. Managers need to be able to design processes, policies and procedures to take advantage of the learning capability provided by the technology. AI technology by itself won’t solve any problems. That is up to us.