AI for Good
Like death and taxes, the inevitable question I cannot escape when on the road is, “Will artificial intelligence (AI) spell human extinction?” We humans have a natural fear of the unknown, so the question is, well, natural. The sentiment comes largely from countless movies portraying robots (with self-awareness AI) turning on their human makers and emotionlessly eliminating them. We fear what we do not understand, and AI (or self-learning) algorithms are far from dinner table conversation for most people. So do we have anything to fear and, if so, when will we need to build that cabin in the Canadian Rockies?
When IBM’s Deep Blue beat Garry Kasparov for the very first time at a game of chess in 1996, and then forfeited their match the following year in 1997, many people thought his defeat spelled the beginning of the end for humanity. While “eliminating humans” might have meant killing them, the same sentiments have tamed a bit, turning to real questions like “Will AI take over the jobs of humans” — and these sorts of questions are relevant and helpful. In the past year alone, at least a dozen headlines in pop-culture and technology magazines alike have asked whether “the robots are coming for our jobs” or have issued cleverly worded and ironic rebuttals to that idea.
Having spent a few decades in the technology industry and having recently founded the first company that uses true artificial intelligence (AI) and machine learning (ML) to prevent cyberattacks — Cylance — I have a different perspective on whether AI and “robots” are going to lead to our doom. We’ve proven the use of AI in real world (both supervised and unsupervised) applications that actually do good, not harm. In fact, I believe that AI has tremendous potential, not only to have a positive impact on the world, but to solve many of the global problems that technology itself has spawned.
From prosthetic hands that employ “touch bionics” and heart valves that can be implanted without invasive surgery to technology that lets people control computers using eye and facial movement, there’s no question that technology can change lives for the better. At the Makerere University School of Computing and IT in Uganda, for example, an AI development group has been using ML to address regional challenges since 2010, applying its platform to work on agricultural blight tracking, traffic monitoring, disease diagnosis, and biosurveillance with good success.
And yet, we know that the creation and operation of the technological devices we use has an impact on the environment. Over a decade ago, a study at United Nations University calculated that producing a computer and its monitor requires 530 pounds of fossil fuels, 48 pounds of chemicals, and 1.5 tons of water. Consider the rapid rate of increase in laptop, tablet, and smartphone adoption within both the business and consumer worlds, and the scale of the impact becomes clear.
That leads us to arguably the biggest problem we share as a planet, which is climate change. Along with a number of concerned global citizens — including Nobel laureates Frank Wilczek and Jose Ramos-Horta, Unilever CEO Paul Polman, and entrepreneur/author Idriss Aberkane — I recently signed a petition calling for millennials to be given a seat at the table of negotiations at the annual Conference of the Parties to the United Nations Convention on Climate Change. In an open letter published in Time magazine, we signers said of today’s youth:
These decision makers, mostly middle-aged and older people, need to recognize that young people have a special role in representing mankind’s interest in our collective future. Most of the people currently deciding how ambitious we should be in reducing carbon emissions and adopting other environmentally sensible measures and policies won’t be around in 30 years.
The generation that will inherit a warmer planet needs the political access and the education to protect that planet. Since I was a young boy, I’ve been taught to help people who can’t help themselves, and I’ve applied that to the protection of our digital lives through AI-fueled cybersecurity. By the same token, I’m interested in helping protect the planet from the consequences of our technological way of life with new ideas that lead to real solutions.
The climate change problem is enormous in scale, and as it turns out, the concept of “scaling” may be key to finding the answer. As the great author and positive-thinking evangelist Norman Vincent Peale once wrote, “Every problem has in it the seeds of its own solution.”
We can apply this maxim to the problem of climate change at two levels: First, in the straightforward sense that we must truly face a problem in order to solve it; and second, with the idea that there are processes found in nature that are mimicked in AI and machine learning that may actually help us tackle global warming.
Idriss Aberkane and I recently wrote about biomimicry, an approach to innovation that looks for long-term solutions to human challenges by emulating patterns found in nature. We explored the idea that biomimicry may save cybersecurity, noting:
Tomorrow’s security software will evolve as a predator-prey dynamic system, with each software population acquiring new characteristics until the entire system stabilizes its software diversity the same way an ecosystem stabilizes its biodiversity…“Natural” intelligence is for the time being still surpassing the one we like to call artificial; in nature, the ultimate measure of intelligence is, well… survival. Deep learning mimics nature: looking at the Internet as an organism, we can attempt to copy the way organisms ensure their internal security.
How poetic would it be if AI could help solve the ecological problem created, in part, by technology in modern life, by emulating strategies, processes and patterns found in nature? In fact, AI/ML is already being used to study climate change:
· Dr. Claire Monteleoni of George Washington University has been using machine learning to study patterns in climate data and to create models that help predict the next stages of climate change.
· The new Center on Artificial Intelligence for Society at the University of Southern California is being driven by the USC School of Social Work and the USC Viterbi School of Engineering. The Center focuses on applying AI to problems including homelessness, health, and climate change.
This is a completely natural application of AI/ML. Not only does machine learning emulate the time-tested survival and problem-solving processes of nature, but it solves the “scale” problem. Tackling our biggest global problems requires the processing of petabytes and petabytes of “big data.” The National Center for Atmospheric Research announced back in 2008 that they plan to preserve and protect 30 petabytes of scientific data over a period of 15 to 20 years.
Machines and machine learning systems are scalable; humans are not. This is why I believe that AI/ML is a natural tool for solving some of the biggest problems on the planet. Will this “tool” be used to take over jobs for humans? The answer is categorically, yes! And it should. We will never be able to solve the really tough problems without it. But can “gut feeling” be programmed into a computer? While this question is a complex one, I believe that eventually almost every aspect of being human can be. Almost anything that we can do as humans, a computer can also do. So what will we do with all the human beings on the planet if everything will be done by machines in the next few decades?
The truth is that the truly difficult and complex problems will always need a human in the mix. The really tough problems always do.
AI is much, much better at learning from past mistakes and changing behavior to remedy those mistakes in the future than humans. AI for the greater good is a concept already being explored and proven within academic and political circles today. At Cylance, we saw that the problem of fighting cyberattacks was a numbers game, with far more threat actors and types of malware than any security team could fight off. Our AI-based approach has turned the tables on that game, allowing security teams around the world to protect invaluable information and our digital ways of life and work.
So I ask you to look beyond the “robot uprising” media memes and consider AI as the problem-solving technology that it is. I believe that our biggest challenges and our greatest hope for global solutions lie in big data and our ability to find patterns and meaning within it. When we leverage AI to solve really hard problems, it will benefit humankind globally.