This beautiful little girl looks like she could be any grade-schooler anywhere. In fact, this person does not exist. It (she?) was entirely created by Artificial Intelligence (AI). The detail in this artificial, machine-created face is astounding. This is possible only since 2014 when Deep Learning, discussed below, was born and computers became “learners” (as opposed to merely “trained”) just like their human overlords.
Artificial intelligence is the most powerful thing that humans have ever created! Although Machines have not yet taken over the world, they are deeply penetrating our lives and affecting our lifestyles. It has the potential to outgrow human intelligence as much as our knowledge surpasses that of snails. That’s really scary for most humans because as the most intelligent species on earth, it’s hard for us even to imagine a world in which we’re not in control. But it is now a possibility for the first time in world history!
Flynn Coleman says, “We are living at the end of the last cycle of technological development led entirely by humans.” (Coleman, 2019) Artificial intelligence is defining and determining the next era of technology and, thereby, our future. AI may possibly be the last frontier of invention and innovation, since our machines will likely become better at inventing and innovating than we could ever be.
Ray Kurzweil said: By the 2040s we will have the opportunity to apply the accumulated knowledge and skills of our civilization to computational platforms that are billions of times more capable than unassisted biological human intelligence. (Kurzweil, 2005) Kurzweil continues saying “The advent of strong AI is the most important transformation this century will see and comparable in importance to the advent of biology itself. It will mean that a creation of biology has finally mastered its own intelligence and discovered means to overcome its limitations.”
Artificial intelligence is showing exponential growth in multiple sectors like medical care, education, voice assistants (e.g., Siri, Cortana, Alexa), facial recognition, natural language processing, self-driving cars and drones, smart thermostats, internet search, military and war-fighting, and many more. Because of its power, AI may become infused in every area of our lives and economy much like electricity. And, incredibly, AI is still regarded as being in its infancy!
However, like every new technology it will be exploited for both phenomenally good and disastrously bad ends. By July 2020, an AI-powered app in the messaging app Telegram had been used to render fully naked renditions of at least 100,000 women as people uploaded innocuous photos of the clothed women for a mere $1.50 per photo. (Hao, A deepfake bot is being used to “undress” underage girls, 2020) The era of deepfakes (fake audio or video produced using artificial intelligence) exploded on the internet in 2018 after a large cache of fake porn videos was discovered that AI had been used to put the real faces of celebrities on the porn actors.
In fact, consistent with the exponential growth of AI, it is being used today to create scores of synthetic humans used to train other AI systems. Because synthetic humans (human faces, human gestures, and human behavior) can be reproduced so accurately using AI, it is easier to use these digital facsimiles to train other AI systems than gather real world sample data about humans. So today, AI is used to generate facial expressions to monitor driver alertness in smart cars, body motions to track customers in cashier-free stores, and irises and hand motions to improve the eye- and hand-tracking capabilities of VR headsets. (Hao, These creepy fake humans herald a new age in AI, 2021) Artificial data has already been used to develop computer-vision systems serving tens of millions of users. In fact, a whole new economy of synthetic data has begun since around 2020 and synthetic data is being mass produced by AI in any number of fields to train, or even, evaluate AI. Financial and demographic data that describes real people can be used to generate detailed fake data set that share the same statistical properties as a company’s real data and simulate data that the company doesn’t yet have, including a more diverse client population or scenarios like fraudulent activity.
AI and Geospatial
Most people haven’t the slightest inkling of how AI in any form functions as a science and as a technology. Yet, understanding the basics of AI and its growing importance to our geospatial businesses is important for at least three reasons:
- AI will impact our businesses both positively and negatively. We must understand and prepare our businesses for these technologies. And,
- AI algorithms are already omnipresent in our daily lives. We are becoming irreversibly reliant on computers and AI machines, using our human memory less, and cognitively offloading more of our tasks, problem-solving, and learning.
- Lastly, these technologies are now advancing at an extraordinarily fast clip. We each must keep up, pay attention, and understand the implications for our businesses, families, and future.
History of AI
So, what is it about artificial intelligence that makes it so important, and making it such a huge deal after only maybe 5 years or so? How is it that AI could be deployed ubiquitously in all areas of human endeavor? To that end, it helps to step back and consider where we are in the larger historical perspective.
The second Industrial Age began about the time electricity was discovered and harnessed. About 100 years later in 1947 the transistor was invented and computers were born. Very quickly this technology grew more powerful, shrunk in size, and declined in cost. Soon, millions of individuals had their own personal computer in the 1980’s. But technology didn’t stop there. Then the world wide web is used to connect all these personal devices to each other in the 1990’s. However, technological progress continues growing at exponential rates and soon (2000’s) everyone on the planet can be tethered to a global network using a tiny personal computer (their cell phones) that are far more powerful than our large desktop computers ever were. But in keeping with Moore’s law, these computers get smaller and cheaper and more powerful.
Then in the early 2010’s we can embed tiny, insanely inexpensive computers into every device. This enables smart thermostats, video doorbells, unmanned aircraft and cars, and sensors in everything and everywhere. Enormous data about everything is being collected by devices and in social media and stored at densities so high and at prices so low that one can scarcely believe it.
All during this period the average cost of storage is dropping consistent with Moore’s Law. Costs went down from $103,000 per Gb in 1990 to $3.50 per Gb today. (McCallum, 2021) Today, cloud computing enables global connectivity of scalable computing that can harness unimaginable power to process this scads of data, and Artificial Intelligence comes into its own in 2017 or so. With AI, machines are not simply programmed, but can be TRAINED to LEARN to extend beyond their programming.
An important concept that explains why AI is important and will become far more important (and possibly dangerous) with each passing day is the idea of “Exponential Evolution or Growth”. The concept of exponential versus arithmetic growth rate is important because all of us living in this technological age have experienced this phenomenon whether we recognize it or not. You see technology can evolve arithmetically or exponentially. Arithmetic growth can be represented as a straight line at some incline from start to finish. But exponential growth says is represented by a curved line shaped like a hockey stick. The “Lily Pond” analogy works well to visualize this concept.
Imagine a small pond with a single lily pad and that the lily pads double in size every day. How long will it take for them to cover pond? The answer is it will take 30 days. The growth and expanse of the lily pads on the lake is exponential. 2 becomes 4 becomes 8 becomes 16 becomes 128 becomes …. 2B very quickly.
But that’s not the interesting part of this story. The interesting part is that the observer standing near the pond doesn’t notice anything until maybe day 23. For 23 days life is “normal”. Nothing appears to change. Then on day 23 our observer may notice that 1% of the pond is covered by lily pads but dismisses it as “no big deal”. Then, on day 29 he takes notice that half the pond is covered. In just 6 days the pond went from “barely noticeable” to “wow!”. Finally, the very next day, day 30, the pond is completely covered. By the time the observer stood up and took notice of the lily pads on day 29, it was too late and BAM! the pond was gone before he had time to react.
Artificial Intelligence technological change is already in the rapidly accelerating part of the curve. We can now “see” the lilies taking over the pond! We NOTICE! Consider the IBM’s computer Watson (in 2011) beat the reigning Jeopardy champion (Ken Jennings) live on TV. Watson digested over 200M pages of data, which it processed at a rate of eighty teraflops (80T) ops per second. But just 3 years late achieved 2400% increase in performance and shrank by 90% from the size of a master bedroom to three stacked pizza boxes! This is happening with Drones, Cars, Lidar, cameras, you name it.
In 2016, The AI system, AlphaGo, defeated the reigning world champion GO player Lee Sedol 4-1 (Rovere, nationalinterest.org, 2020). Google trained AlphaGo with 1000’s of hours of game play, taught it all the rules of GO. This AI system learned, and learned, and finally beat the world champion.
But that was nothing: two years later, a new AI system, AlphaZero, was pitched against AlphaGo. Unlike its predecessor which contained significant databases of go theory, all AlphaZero knew was the rules, from which it played itself continuously over forty days. After this period of self-learning, AlphaZero annihilated AlphaGo, not 4-1, but 100-0 (Rovere, nationalinterest.org, 2020). In forty days AlphaZero had superseded 2,500 years of total human accumulated knowledge and even invented a range of strategies that had never been discovered before in history.
As important as this concept is there is yet another important idea we need to understand that affects our understanding and perception of technological change especially as it relates to AI.