The Exponential Guide to Artificial Intelligence
“AI is here today; it’s not just the future of technology. It’s embedded in the fabric of your everyday life.” — Neil Jacobstein , Singularity University Chair, AI & Robotics
Today, it can be difficult to understand the significance and potential impact that artificial intelligence (AI) has for humanity. From Siri to IBM’s Watson to Hollywood portrayals of killer robots, it’s not clear what we should ultimately expect from this exponential technology.
What is clear: AI-powered products and services have made it into nearly every aspect of our personal and professional lives in just a few years. And as AI solutions continue to emerge and converge, that pace of change will only continue to accelerate. It’s easy to find scenarios of a utopian future of abundance where machines do all the hard work—as well as grim scenarios where unemployment soars as traditional workers are replaced by increasingly capable machines.
With such rapid progress, it’s difficult to make assumptions about the future of AI. But instead of focusing on the unknown, we can examine what we know about AI, its current applications, and potential future impact.
At Singularity University, we help organizations and individuals understand the disruptions and opportunities of exponential technologies like AI. Whether you’re an entrepreneur, Fortune 500 CEO, or simply a curious human who wants to understand where we’re going as a species, AI is significantly impacting all of our lives.
None of us can predict the future of AI. But if you’re looking for an accessible guide to help you understand this exciting technology, we offer this Exponential Guide to Artificial Intelligence. Read on for more!
What Is Artificial Intelligence?
AI is an “umbrella term” for a branch of computer science focused on creating machines capable of thinking and learning. Based on their experiences, AIs learn to make better decisions in the future. This ability to both learn and apply knowledge closely mimics the way human beings understand the world and allows machines to accomplish tasks that were once only possible with human minds.
Some of the human-like tasks AIs can do include:
Complex problem solving
Visual interpretation (computer vision)
Speech recognition (natural language processing)
These capabilities are accomplished via a collection of computer algorithms that use mathematics and logic to perform the AI’s assigned task. So although our most famous science fiction books and movies tend to portray AI in the form of human-like robots, AI is simply computer code running in software.
Unlike the human brain, these intelligent programs can be run in a variety of different hardware types, whether that’s your smartphone, a warehouse of web servers, or a self-driving Tesla.
This variety of use cases is what often makes AI so difficult to understand, but it’s also what makes it so powerful. The ability to add an AI layer on to nearly every technology means that as AI progresses, the world around us will increasingly seem to come alive. This “awakening” will drastically alter life as we know it, from leisure and business activities to our health and spirituality. To get an idea of how this might happen, let’s first take a look at how AI works.
“AI is perhaps the granddaddy of all exponential technologies—sure to transform the world and the human race in ways that we can barely wrap our heads around.”
How Does Artificial Intelligence Work?
Much like human intelligence, AI works by taking in large amounts of data, processing it through algorithms that have been adjusted by past experiences, and using the patterns found within that data to improve decision-making.
To simulate human intelligence in this way, AI engineers provide their machines with ability to:
Perceive their surrounding environment (which may simply be data)
Detect patterns in the environment
Learn from the patterns and update experiential memory
Then, these steps are repeated until there’s enough data to confidently make predictions and support decision-making.
What makes AI remarkable is the speed, accuracy, and endurance it brings to this human-like learning process. Humans have to eat, sleep, and tend to a variety of personal needs. We are also creatures of comfort, and quite stubborn—too much change makes us uncomfortable. And when presented with new information and experiences, humans tend to let our biases sway us from making the most reasonable and logical decisions.
Machines suffer from none of these shortcomings. For most purposes, they’re capable of running indefinitely, allowing AIs to process and detect patterns in massive amounts of data without mental fatigue.
AIs are constantly tweaking their understanding of their environment, updating their “perspective” of reality, and updating the probability of their predictions without clinging to any old ideas. Some people find this cold logic the most terrifying part of AI, however, it’s also what allows AIs to find solutions humans may not recognize.
The concept of AI has been around since 1955, but its growth has exploded in recent years because of three factors:
Vastly increased computing power
Large, inexpensive data sets
Advancements in the field of machine learning
But computing power alone wouldn’t have accomplished much if not for two key technologies that support AI: big data and machine learning.
Big data, which provides massive data sets and user activity to greatly increase the quality of “education” AIs receive.
Machine learning is a method of data analysis that enables computers to learn without external instruction.
Deep learning is a branch of machine learning that uses computer simulations called artificial neural networks.
How Are AI, Big Data, Machine Learning, and Deep Learning Related?
As we’ve mentioned, AI covers a broad field of sciences involved in developing computer systems that think and learn in a way that’s similar to human intelligence. AI applications are often divided into “narrow AIs” that perform specific tasks such as playing chess, and “general AIs” that understand language, context, and emotions as humans do. Let’s take a closer look at the relationships between AI, big data, machine learning, and deep learning.
With the rapidly decreasing cost of sensors and the global growth of the Internet of Things (IoT), we have dramatically increased the number of smart and connected devices that are continuously measuring and recording data. Nearly every action we take is now recorded in a database somewhere. This includes mobile device activity, the purchase history on our credit cards, our online browsing activity, our social media feeds, and even our biological data.
Big data is the term for these massive collections of data that we’re all contributing to every day. Big data is the fuel that enables AIs to learn much more quickly. The abundance of data we collect supplies our AIs with the examples they need to identify differences, increase pattern recognition capabilities, and to discern the fine details within the patterns.
If you provided an AI with one picture of a dog and one picture of a cat to learn from, you will have an AI that’s terrible at the task of determining pet species. Feed that same algorithm millions of pet pictures, and the AI can quickly learn how to distinguish dogs from cats, and also determine the different breeds within the species.
Big data enables AIs to learn by example rather than by instructions provided by humans. And they’re able to learn this way because of the advances in machine learning.
Machine learning is a method of data analysis that learns from experience, enabling computers to find hidden insights without being explicitly programmed to do so. Machine learning analyzes data and learns from it to make decisions and predictions, and includes supervised (manual entry of data and solutions) and unsupervised learning.
Machine learning is a subset of the larger field of AI, and it is one of the many processes that enable the creation of AI. Many ways of creating AIs have been explored, but machine learning is important because it does not require human input or interaction. Rather than learning by instruction, machine learning AIs learn by exposure to examples found in data. Through machine learning, AI is able to take advantage of the enormous data sets generated by our daily activities. To learn without human involvement, machine learning works largely by implementing statistical methods into the learning process.
Deep learning is part of the broader field of machine learning that uses artificial neural networks, which are computer simulations patterned after a human brain. Deep learning includes aspects of machine learning algorithms, neural networks, and AI.
The artificial neural networks created from these components are where the field of AI comes closest to modeling the workings of the human brain. Improved mathematical formulas and increased computer processing power are enabling the development of more sophisticated deep learning applications than ever before. Deep learning—also called structured learning and hierarchical learning—is the kind of machine intelligence used to create AIs that beat humans at games of Go and chess.
Watch Singularity University Co-Founder and Chancellor Ray Kurzweil talk about deep learning and the path to artificial general intelligence in this captivating video.
How Does AI Affect Our Lives?
Some of the most powerful and prevalent applications of AI are the ones we often take for granted. These include the AIs that handle your Google searches, deflect spam from your inbox, and select the ads you see across the digital landscape. AIs identify people in your Facebook pictures, and recommend the products you buy from Amazon.
No matter where you live and work, one thing is certain: more and more of our society’s technical infrastructure is powered by AI. While many AIs are easy to overlook because they don’t talk to us like Siri or perform physical tasks like driving our Teslas, they constantly work behind the scenes, performing crucial functions like pattern recognition, problem solving, reporting, and optimization.
AI technology is making its way into nearly aspect of our lives. It’s helping to keep us alive through its integration with healthcare, and influencing our economies via its integration with finance.
Learn more about the powerful potential of AI in medicine at Singularity Hub and join us at Exponential Medicine in San Diego, California in November 2019.
What Are Some Examples of How AI Is Impacting Healthcare?
With the fundamental importance of health in our lives, it should be no surprise that we’re seeing a massive integration of AI throughout healthcare and medicine, from cybersecurity for patient records to AI-assisted surgeries. Here are some examples:
One study showed how virtual assistants running natural language AI systems are saving doctors and nurses 17-20 percent of the time by cutting back on unnecessary visits and workflow overhead.
New AI implementations are being used to discover gaps in patient care, protecting against oversights for scheduling and treatments, which helps hospitals improve care and potentially prevent malpractice lawsuits. It has been estimated that using AI to streamline general administrative workflow at hospitals might provide an annual $18 billion in savings .
Diagnostic practices are benefiting from the ability of AIs to quickly and accurately analyze samples.
In pharmaceutical research, AI is being used to massively speed up the process of drug discovery.
From helping human healthcare employees work more efficiently, to improving diagnoses and discovering new drugs, AI stands to revolutionize an industry that, became the largest U.S. employer in 2017 .
The majority of women treated for late-stage breast cancer receive the wrong treatment in the first year because the only way to see if one of 30 FDA-approved drugs will work is for the patient to try it to see what happens.
Ourotech, a Singularity University Portfolio Company, is doing something about it. Learn about a major breakthrough that has led to a revolutionary way to treat late-stage breast cancer, thanks to AI.
Read the case study
What Are Some Examples of How AI Is Impacting Financial Services?
The strengths of AI are a good match for the challenges facing financial services firms around the world. AI has generated a lot of excitement and attention in recent years because of its huge potential to add value to all kinds of financial services transactions. Banks and investment firms are exploring the power of AI to improve customer experience, automate cumbersome tasks, cut costs, and help uncover new opportunities for future growth.
For example, the ability of AI to detect and analyze patterns in big data makes it a powerful tool for wealth management and investments. One of the key ways we’re seeing this partnership today is via AI-powered “roboadvisors” that are taking on many aspects of financial portfolio management for clients.
Companies like Betterment that use a combination of human and AI expertise are leading the charge in this growing trend. The company helps customers set up a portfolio, choose, and maintain investments for a fixed annual fee. Betterment’s approach has gained popularity in recent years and the company currently oversees more than $10 billion in assets for over 250,000 customers.
And for those of us who are concerned with the security of our personal bank accounts and assets, we can expect more sophisticated, AI-powered fraud protection in the future. And for those of us who have endured cumbersome and unhelpful phone support from our banks, we can look forward to advances in AI service bots that promise to be much more efficient at problem-solving and providing quick responses.
What Are the Risks and Benefits Associated with AI?
“People are really too focused on ‘evil AI,’ and not focused enough on human intent.“
There is a popular argument that tools like AI essentially are neutral, and can be used for good or evil, depending on the user’s intentions. While AI is unique in that we’re building it to be capable of developing its own learning and “intentions,” it’s realistic to expect that for the foreseeable future, AI will be shaped by the direction of its human creators.
We can say with certainty that AI is such a profound tool that its impact marks a true global paradigm shift, similar to the revolutions brought about by the development of agriculture, writing, and manufacturing.
While the future changes that AI will bring are almost impossible to imagine, we have identified three key benefits and three key risks worth keeping in mind:
Risks of AI
Drastic changes to our lives
AI created with bad intention
AI created with good intention goes bad
Benefits of AI
Liberate humans to do what they do best
What Are the Benefits of AI, in Greater Detail?
In an ideal world, AI represents a win-win scenario by providing strengths that humans don’t possess. Advanced pattern recognition, computing speed, and nonstop productivity courtesy of AI allow humans to increase efficiency and offload mundane tasks—and potentially solve problems that have evaded human insight for thousands of years. Let’s look at some benefits of AI in more detail.
AI offers increased efficiency
We are human, and so we make mistakes and get tired. We can only perform competent work for a limited time before fatigue takes over and our focus and accuracy deteriorate. We require time to unplug, unwind, and sleep.
AIs have no biological body, side-gig, or family to pull their attention away from work. And while humans struggle to keep focus after a while, AIs stay as accurate whether they work one hour or 1,000 hours. While they work, these AIs can also be accurately recording data that will, in turn, provide more fuel for their own learning and pattern recognition.
For this reason, AI is transforming every industry. The amount of time and energy companies have to invest in repetitive manual work will diminish exponentially, freeing up time and money, which in turn allows for more research and more breakthroughs for each industry.
Keep exploring: Learn how to build an enterprise AI capability in eight steps.
AI is solving problems for humanity
As AIs gain greater capabilities and are deployed in different capacities, we can expect to see many of the problems that have plagued government, schools, and corporations to be solved. AIs will also be able to help improve our justice system, healthcare, social issues, economy, governance, and other aspects of our society.
These critical systems are rife with challenges, bottlenecks, and outright failures. In each realm, human bureaucracy and unpredictability seem to slow down and sometimes even break the system. When AIs gain traction in these important domains, we can expect much more rational, fair, and thorough examinations of data, and improved policy decisions should soon follow.
AI is liberating humans to do what they do best
As AIs become more mainstream and take over mundane and menial tasks, humans will be freed up to do what they do best—to think critically and creatively and to imagine new possibilities. It’s likely this critical thought and creativity will be augmented and improved by AI tools. In the future, more emphasis will be placed on co-working situations in which tasks are divided between humans and AIs, according to their abilities and strengths.
Perhaps the most important task humans will focus on is creating meaningful relationships and connections. As AIs manage more and more technical tasks, we may see a higher value placed on uniquely human traits like kindness, compassion, empathy, and understanding.
What Are the Risks of AI, in Greater Detail?
Will AI change our current way of life? Absolutely. Do we know exactly how? Absolutely not.
AI already is affecting nearly every aspect of our personal and professional lives. Every human institution—businesses, governments, academia, and non-profits—is already experiencing the accelerating pace of change. And although AI is often portrayed in terms of solutions to solve problems in healthcare, transportation, and business productivity, there is also a darker side to consider.
There are concerns that AI will replace human workers, and some people fear the ultimate outcome will be that superintelligent AI-powered machines will eventually replace humans entirely. While this is a possibility, many experts believe that it’s more likely that AIs will enhance, not replace, humanity and that eventually, we might merge with AIs .
It’s essential to think about what might happen when a tool as powerful as AI malfunctions or is used with malicious intent. Consider the following two scenarios:
Scenario 1: AI created with bad intentions
Those who insist that technology is neutral will point out that a hammer can be used to build a home or to hit someone over the head. As with any technology in the wrong hands, AI could be created to help humans commit horrible acts. This might be an autonomous weapon programmed by the military, or a malevolent algorithm set loose by an individual hacker.
Fear associated with AI—a technology that is intelligent and capable of self-learning—is not unfounded. But it’s important to remember that humans also are highly intelligent and capable of rapid learning and improvement.
Moreover, it’s also worth remembering that harmful AI capabilities aren’t created in a vacuum. While one person or group is attempting to create something harmful, there is often an equal or greater amount of energy being invested to stop that harm and create countermeasures that limit risk and impact.
Scenario 2: AI created with good intentions goes bad
Another scenario is the runaway AI, in which a machine that was built with good intentions turns bad—a staple of classic Sci-Fi films like “Blade Runner” and “2001 Space Odyssey.” Indeed, when the sentient computer HAL turned against astronauts in the 1968 Stanley Kubrick film, many viewers found the premise to be unrealistic. With the widespread use of AI, as well as its growing capabilities, this scenario may no longer seem as far-fetched.
Addressing concerns over whether AI will drive massive job displacement , Singularity University Co-Founder and Chancellor Ray Kurzweil explains that while certain jobs will be lost, new jobs and careers will be created as we build new capabilities.
Kurzweil notes that AI will benefit humans and that AI is less likely to be threatening than beneficial to us , and it benefits us in many ways already. In Kurzweil’s view, a robot takeover is less likely than a co-existence, where machines reinforce human abilities and accelerate our progress.
Resources for Additional Learning
What Are Some Leading Trends in AI?
As the development and application of AI continues to evolve at an unprecedented rate, a handful of important trends have begun to emerge. Perhaps the most significant trends involve deep learning applications that have demonstrated outstanding performance competing against human contestants in games like Jeopardy and Go. The job market also reflects this growth clearly. From 2015 to 2017, for example, we saw a 35x increase in posted jobs that require deep learning development skills . And in 2019, the demand continues to increase.
One reason for AI’s powerful growth is its convergence with other technologies. We’re seeing a massive increase in AIs’ integration with the Internet of Things (IoT), and with edge computing, a strategy designed to increase performance by moving computing power out of data centers and closer to local devices. The purpose is to enable devices to respond faster by processing more information locally, rather than sending the communications back and forth to the cloud. The integration of AI, the IoT and edge computing will be a driving force as businesses seek to improve the speed and performance of their solutions and services.
Another important trend is the development of specialized processors that are engineered to optimize AI performance. Some of the world’s premier chip manufacturers, including Nvidia, Intel, AMD, Qualcomm, and ARM are all working on their own versions of high-performance chips that will enable AI’s deep integration into everyday products and the IoT.
Other important trends driving the growth of AI include computer vision, voice assistants, and a push for more standardization and ethics.
AI Is All Around Us