Читать книгу Outsmarting AI - Brennan Pursell - Страница 3

Sean Kanuck, former US National Intelligence Officer for Cyber Issues

Everyone needs to grasp the impact of emerging technologies on their lives, organizations, and societies in order to survive and prosper. The impending combination of information technology (IT), nanotechnology, and biotechnology will literally change life as we know it—in almost every respect. The scale and scope of that change will be monumental. Not since mankind gained command over fire have humans been in an equally powerful position to transform our own existence. The ability to cook food altered our bodies, and the ability to forge metal defined early cultures. But, at the same time, certain aspects of our identity and experience remained constant (e.g., reproduction, consciousness, mortality). The same will be true for the foreseeable future of this technological revolution.

Moreover, the exigencies of climate change and global pandemics will only accelerate the pace of change. The systemic sources and adverse consequences of increasingly violent weather patterns, ecological disruptions of flora and fauna, novel diseases, drought, and famine are undeniably linked to one another. People must develop a better understanding of how our planet works—as well as how our own bodies work—to learn how we can better manage resources in a sustainable fashion for public health and welfare. That will require diagnostic sensing and data processing at hitherto unprecedented levels.

The great IT advances being discussed today include the Internet of Things (IoT), blockchain, quantum computing, and artificial intelligence (AI). Among those, AI perhaps holds the most promise and is also often the most misunderstood. I am honored and privileged to write this foreword for a book that endeavors to simply and accurately convey both the current state and future potential of AI for nontechnical audiences. This book serves an important purpose in making complex issues accessible and understandable. It also places AI in the appropriate context for a number of business scenarios and applications. By sharing these important insights, it will help empower a wide range of enterprises to benefit from those innovations.

As a strategic analyst and business consultant, I offer that AI will be in an elite class of technologies that truly live up to the hype—but that will take time. I concur with those who claim that AI will be as transformative as the harnessing of electricity, and moreover, that we cannot even begin to predict the multifarious applications that it will enable. For who could have foretold Benjamin Franklin, Alessandro Volta, Michael Faraday, Georg Ohm, or Thomas Edison the uses to which modern society would put their discoveries? But, while we may lack specificity about the future, we can recognize certain principles and begin designing frameworks to leverage new capabilities. For corporate executives, investors, and entrepreneurs alike, that means preparing for the marketplace of tomorrow. By way of example, we are already seeing dramatic changes to teleworking, logistical supply chains, and public health procedures as a result of applying IT solutions, including AI, to enable social distancing, home delivery, and contact tracing in the post-COVID business world.

The very first step in capitalizing on the potential of AI is to realize that it is not just a technology issue. How enterprises purchase and employ it will soon be a central business strategy issue. We do not talk about companies that do or do not have electricity; we talk about how they utilize it and how much they have to pay for it. The same is already true of the Internet (including cloud computing resources), and the same will also soon be true of AI. AI is more than a topic for the chief technology officer; it is a compelling issue for the entire “C-Suite” (i.e., chief executive officer, chief strategy officer, chief product officer, chief financial officer, chief risk officer, etc.) and every board of directors.

The second framing concept I would like to offer is that all data are not created equal. AI algorithms require troves of data to operate, and in essence, they are what they eat. In other words, the quality (e.g., validity, utility, etc.) of an AI algorithm’s outputs will be dependent on the quality (e.g., volume, accuracy, etc.) of its inputs—including during the training phase of machine learning. Not all raw data should be viewed as useful information. In many cases, it needs to be structured, formatted, and checked for accuracy before it is fed into an algorithm. For that reason, I have likened data to seawater: More than 70 percent of planet Earth is covered by oceans, but we must expend time and energy to purify that water before it can be used to quench our thirst or hydrate our bodies. Where your organization obtains and how it handles its data will matter deeply.

Those first two principles of AI management converge on a third issue of utmost importance: law and ethics. Once we realize that AI algorithms harbor the innate biases of their programmers and can generate (and perpetuate) erroneous conclusions based on flawed data sets, we must also acknowledge a social responsibility to minimize those errors. Racial profiling and misogyny are two adverse factors that have already been proven to propagate among ill-managed AI programs. General counsels will have an interest in limiting an organization’s possible legal liability, while human resource managers will have a desire to fully capitalize on the potential of a diverse workforce. In this case, doing what is right will also help the bottom line.

So, as you read this volume by Brennan Pursell and Joshua Walker, keep those three notions at the forefront of your mind. Their chapters that follow will provide you with many useful ideas for incorporating AI into productive business models, but they will also identify the limitations and potential shortcomings of the technology itself. AI is not a panacea; it is yet another human-engineered tool to aid us in conducting (previously) human tasks at much greater speed, scale, and accuracy.

In parting, let me offer what I believe will—and will not—change in the marketplace of the future. For starters, businesses that embrace AI will far outpace their competitors. There will be “haves” and “have-nots,” and the disparity between the performances of those two genres will become acute and pervasive. Enterprises that learn to leverage data and employ AI algorithms effectively will have a comparative advantage because they will operate more efficiently and be able to perform functions that surpass human sensory and computational capabilities. An unfortunate corollary of the fact that all data are not created equally, however, is that large organizations with myriad data will have an inherent advantage over smaller competitors. Consequently, my advice would be to purchase access to AI algorithms trained on those larger data sets.

Next, AI—in conjunction with robotics—will displace workers in multiple roles, including some professional and managerial jobs. But, these new applications will themselves require human facilitators and programmers to calibrate their activities. Once again, the issue will be how successfully an organization manages the inevitable change to its own advantage.

Finally, certain elements of corporate competition and responsibility will remain. The transition to an AI economy will see many established companies flounder, and many new enterprises supplant them. But the age of the truly intelligent machines is still a matter for science fiction. Human beings will continue to matter, and companies that respect the human dignity of both their employees and their customers will increasingly see that reflected in their profits.

Sean Kanuck is chair of the Research Advisory Group for the Global Commission on the Stability of Cyberspace. He teaches a graduate seminar on the security implications of artificial intelligence at George Washington University’s Elliott School of International Affairs.