Machine intelligence (MI)—the ability of machines to perform tasks that would normally require human intelligence—is becoming an increasingly urgent topic of discussion in c-suites, newsrooms, and academic institutions around the world. Rapid gains in computing power, exponential increases in availability of digital data, and new research in computer and data science are enabling algorithms to meet or exceed human capability across diverse tasks. These developments have given rise to machine capabilities that at once inspire and unnerve: trucks that drive themselves, computer programs that develop drug therapies, software that writes news articles and composes music.
Many executives find themselves alternately exhilarated, unsettled, and bemused by the predictions about machine intelligence’s (MI) potential following its recent technical advancements.
Self-described “tech evangelists” predict MI will soon slash the cost of doing business, add billions or trillions of dollars to the economy, and liberate employees from drudge work. Cynics caution with equal vigor that we are on the verge of upending financial markets, seeding mass unemployment, and perhaps threatening humanity itself (much good cost savings will do us then). A growing cadre of skeptics disavow both positions, pointing to all the times so-called “artificial intelligence” has failed to deliver in the past and continues to now. Who hasn’t cursed a virtual customer service agent?
So, it would be tempting for leaders to set MI exploration aside until there is greater consensus about its capabilities, benefits, and risks. But waiting would be a mistake. MI, to be sure, is a nascent technology with serious limitations. It is often difficult to implement and falls far short of our expansive human capabilities. But even in its current, limited form, MI is being used to substi- tute for human intelligence in a growing number of tasks. These tasks, taken together, account for meaningful portions of our professional and personal lives. Ceding them to machines—as we’ve just begun to do—will change our jobs, economy, relationships, and understanding of ourselves. The gravity of this impact calls us to act. We must create—not respond to—our future with MI. We will decide what MI means for our collective future.
The first step is cultivating our understanding of what MI is and what it implies. This Machine Intelligence Primer provides this foundational understanding. In it, we discuss where MI came from, how it got to where it is today, and where it’s likely going. We explain and dispel common myths that surround it. While the Primer discusses technical topics, no technological expertise is required to understand it. Instead, it is meant to help executives, practitioners, and curious skeptics alike consider what MI will mean for them and their teams. We hope it will help you consider how MI might be used to create a world that is not simply more efficient, but more equitable, meaningful, and verdant.
In creating the Primer, we were inspired by the works of wide-ranging academics, entrepreneurs, and government organizations whose work came before ours. We hope this book will inspire you to watch, read, and listen to the insights of MI experts like Andrew Ng, Stuart Russell, Fei-Fei Li, and John Launchbury. We are honored to accompany you in your effort to understand and harness MI. We hope that you will share the ideas, aspirations, and worries the information in the Primer evoke for you. We are interested to hear these perspectives from you as we set out on our own journey to make the most of MI.
Please join the conversation and contact us at machineintelligence@bah.com.
