Competing in the Age of A.I. — Interview with Marco Iansiti and Karim R. Lakhani
In this episode of Masters of Data, I speak with Harvard Business School professors Marco Iansiti and Karim Lakhani. We discuss AI and its increasingly important role in business and the economy. After studying digital transformation for the last ten years, Marco and Karim noticed a distinct change in the way the economy works and found similar themes emerging in their research. As a result, they recently published Competing In the Age of AI, a book that discusses how companies can use AI to be more competitive, and how data and analytics are more important than ever to the success of a business.
For the last few centuries, firms have been all about people, managers, and workers, and driving productivity gains for workers. In the last few decades, in particular, information technology has been sprinkled in to make these workers more productive. In fact, in today’s economy technology is put front and center and expected to deliver success for the business and value to customers. You often hear that such and such a business is really a technology business, as opposed to say a manufacturing or retail business. For example, while Uber is a part of the transportation business, Uber’s software is essential to its success. But this really oversimplifies the roots of success for today’s powerhouses like Google, Microsoft, and Facebook. These companies have changed their business models to rely not just on software, but more specifically on Artificial Intelligence (AI) and data to drive exponential growth. Software may be “eating the world”, but now AI, powered by data, is eating software.
Marco and Karim present three elements to consider for companies who are attempting to build operating models around A.I.: scale, scope, and learning. They give the example of algorithms that determine whether or not someone will get a loan — a process, using A.I., that basically costs a company nothing. By using A.I. to remove humans from the process, a bank can scale this function orders of magnitude beyond traditional methods. Next, with an A.I.-driven operating model a company’s scope has also grown; now, you can plug things into a software core and use the data gathered from one application to drive value to the other. Lastly, A.I. has increased the opportunities for businesses to learn. With this new operating model, the more people a company serves, the more data they gather. And A.I. and Machine Learning improve with the greater access to data, thus increasing the scale and scope of the business — a virtuous cycle.
With this new A.I. driven operating model, there is no human error, and it frees up employees to do other tasks. AI is not replacing humans, but simply moving them to the edges. Humans still add value, but they are no longer in the critical path of execution limiting scale. These three concepts — scale, scope, and learning — define the operating model of a company.
AI isn’t new, Marco and Karim point out; it is just finally coming into its own. A business cannot simply place AI into existing processes. Instead, Marco and Karim suggest that it is necessary to redesign processes to take advantage of the technology. Businesses that are willing to take the time to incorporate A.I. will survive much longer than those that won’t. This transformation often takes a long time and changes the way organization function — as well as the organization’s culture, value systems, and structure. The concept of an AI factory, a common set of A.I. infrastructure, algorithms, and data, is the beating of this new model. An AI factory’s key input is data, something which is typically undervalued and often massively uninvested in. And the A.I. factory can be applied to all industries. For example, though McDonalds’ hamburger factory looks very different from Ford’s steel factory, the two companies’ AI factories look exactly the same.
One important point raised by the authors was that while algorithms are not always complicated, but they should be practical and provide value. Once you replace even a simple process that humans are bottlenecking with an algorithm, the results can be dramatic. Algorithms, Marco and Karim reiterate, save time and create accuracy. However, algorithms are very narrow and do one thing — such as recognizing images, noticing tumors, playing chess — very adeptly. The most powerful AI is a collection of these narrow algorithms that work together to something more than the sum of its parts. People often wonder when AI will become smarter than humans, and Marco and Karim respond with, “How do you know we’re not there yet?” The Google search engine is certainly smarter than us, but in a different way. AI is, again, a combination of “dumb” algorithms working together to do something incredible. Therefore, they say, AI won’t replace humans, because AI is simply smart in different ways. When asked this question, Pedro Domingos, a science professor at the University of Washington, said that AI will not replace managers, but managers with AI will replace managers without it.
As part of the digital transformation, companies still need wise leaders. Wisdom is something that is often lacking in leadership, and Marco and Karim believe that wisdom is essential to success. There is ongoing friction between ethics and technology because every human being is biased in one way or another, whether consciously or consciously. The programmers writing the code for the AI are biased, and those overseeing the AI are biased. There is a need for engineering leadership to consider ethics from the ground up. It is also very important for C-Level executives and other leaders to take a deep interest in the work of AI and its role in their company. Top-down support is necessary for success; the leaders need to understand their systems so that they can make informed decisions, and their support of AI signals to the rest of the organization that AI is important.
Marco and Karim end the episode by asserting that while some companies are struggling to make this transition, the economy cannot afford to lose its best firms, such as Ford, GM, and GE. The men both work in a lab at Harvard that conducts research on the basic challenges of digital transformation. They are studying the value of data and its competitive advantage and running projects that assess how advanced companies operate, and where the world stands in relation to AI and information technology. Both Marco and Karim encourage listeners to reach out to them if they have any case studies, examples, or any other helpful observations to share.
Outbound Links & Resources Mentioned
- Listen to this episode of Masters of Data
- For more on Masters of Data, head over to http://mastersofdata.com
- Purchase Marco and Karim’s book, Competing In The Age of AI
- Connect with Marco Iansiti on LinkedIn
- Read more about Marco on his faculty page
- Connect with Karim R. Lakhni on LinkedIn
- Read more about Karim on his faculty page
- Learn more about Harvard Business School
- After ten years of looking at digital transformation, Marco and Karim noticed a huge shift in the way AI affects firms, which spurred the writing of their book.
- For hundreds of years, firms have been all about people, and now they are increasingly about technology.
- Most businesses are now implementing software, making more and more businesses software businesses.
- Companies apply technology to solve a business problem and give them an edge on their competition.
- The three challenges that Marco and Karim present are scale, scope, and learning.
- Companies can use algorithms for new levels of scalability.
- The more data a company gathers, the more opportunities they have to learn.
- AI is not replacing humans, but moving humans to the edges. Humans become supervisors of AI.
- Businesses have to redesign their processes to take advantage of new technologies.
- AI factories look exactly the same across companies, while their physical factories look much different.
- Algorithms are usually very basic and narrow, but when working together, they can render incredible results.
- AI is not necessarily smarter than humans, but the way AI approaches problems is just much different.
- The problem of technology implementation is not based on how old a company is; there are still modern companies that have not been able to tackle it.
- There is ongoing friction between ethics and technology because of unconscious bias.
- It is important for CEOs and company leadership to take a deep interest in the work of integrating AI.
- Marco and Karim have a laboratory at Harvard that researches many topics related to AI and information technology.