WHAT IS AI? EVERYTHING TO KNOW ABOUT MACHINE LEARNING
When people hear “AI" they often think about robots overtaking the world like in Hollywood movies and TV shows we have come to love, and be afraid of. But how soon will AI (Artificial Intelligence) as we’re imagining, actually become a reality? To know how soon those fantasies will be our lives, we decided to learn the history of AI, and where mankind is today with the development.
The short history of AI really began after World War II. An English mathematician, Alan Turing, hypothesized if computers could think like humans. He presented a lecture in 1947 outlining how AI development would come from computer programming rather than mechanical machine development. By the 1950s, AI was popular among researchers world-wide.
In the 1960s, the US Department of Defense took interest in AI. Long before our home “assistants” like Alexa or Google, the government was programming problem-solving algorithms, organizing data, and mapping communication by-way of computers. These were the first steps in creating what we know as the home computer, software, and cloud capacities we use every day today.
Then in 1996, IBM’s computer, Deep Blue, won a game of chess against the current World Champion of Chess player, Gary Kasparov. It was an essential showcasing of computers learning and beating humankind at their own development. And was just the beginning of a race to make machines more intelligent than humans.
So how would we define AI as it’s used today? In short, AI is the computer science concerned with building smart machines that can learn from themselves exponentially faster than humans. This affects our lives through almost every industry. Everything from airplane flights, to online shopping, to material logistics, to healthcare can become more effective. And computers can optimize these industries exponentially faster than humans. Artificial Intelligence is creating these advancements.
A second definition of AI, as written on builtin.com is, “the endeavor to replicate or simulate human intelligence in machines.” This meaning might hit a bit closer to sci-fi books and theatrical fantasies. But that’s frankly not all of AI. Generally speaking, AI can be simple, “Narrow” or “Weak AI” where the computer does one task very well. All the work Alan Turing, and the USDOD completed led us to sophisticated uses of this. An example today would be Google search. The program calculates algorithms to list your search results in the most effective manner. It’s able to search and prioritize considerably faster than a human could. But it has really optimized this single ability. A second form of AI, “General Intelligence” or “Strong AI” is more like robots we imagine - where computers are mimicking human behavior and mindset, encompassing and judging and solving much more complicated tasks simultaneously. Some experts note that we’re far from Artificially Intelligent robots (or AGI) going to war with humankind. Legendary investor Naval Ravikant explains, “to actually model general intelligence you run into all kinds of problems. First, we don’t know how the brain works. At all. Number two, we’ve never modeled a paramecium or an amoeba let alone a human brain.” Naval continues to explain how so much of the brain work is happening within a cellular level - more microscopic than the cellular level itself, and we’re nowhere near replicating either level of calculations or understanding the environment to a point where a replication success can exist. So creating an AGI robot, though the goal of lots of AI developers, is far from our modern day reality. Until then, we continue to develop AI to replicate human intelligence as close as possible.
Still, there are current-day pros and cons to AI development. A certain type of AI, “machine learning” is when computers are learning algorithms, maybe thousands at a time, and perfecting them to be as close to ideal as possible. So, for example, when you’re shopping online and the website provides options not just suitable for your taste, but exactly what you want. The retailer has used AI machine learning to study your wants and needs, then show the exact solutions to you. This can be really positive AI. Or AI allows us to get real-time directions with a couple taps to our phones. Another great advantage AI has brought to us. In healthcare AI can detect cancers, heart disease, and even schizophrenia, sometimes even more accurately than human doctors. Logistics around the world can improve too. Everything from tracking inventory, to tracking trucks, predicting demand, to answering the most requested customer service questions can be optimized using pattern recognition and AI. Overall, AI development is making our lives more streamlined, painless, and smooth. In fact, in 2016, President Obama’s CTO (Chief Technology Officer) Megan Smith signed her name to a White House AI briefing that said, “Although prudence dictates some attention to the possibility that harmful super- intelligence might someday become possible, these concerns should not be the main driver of public policy for AI.” It’s good to know Megan Smith had worked for years at Google, leading narrow AI teams for the private corporation. Still, as of 2016, the most powerful office in the world was being told not to worry about AI.
However, there are those that worry about our rapid AI development. Elon Musk famously invested in Google’s DeepMind London based research lab, then seemed to learn too much. “AI could surpass human intelligence in the next five years, even if we don't see the impact of it immediately.” He said in 2020. The fact that computers can now learn extraordinary complicated patterns and map them faster than humans, is concerning. Intelligent machines can now master chess in 9 hours, and the very strategically complex Chinese game of Go in 34 hours. In fact, the human world leader of Go, Lee Se-dol, retired in 2016 after having been beaten by AI. What does this mean to us? Computers can learn infinitely faster than us. And they can learn from their own examples without having human examples demonstrated for them. They adapt and strategize very quickly. Sometimes simply the number of patterns they are solving, is incomprehensible. If these computers were to analyze, define, and map war tactics this can be extremely effective and mean a low death count. But giving the computers the ability to do this runs along the edge of becoming dangerous the moment the computer strays from human control. And our technology developers today don’t necessarily have a great track record of knowing where that line of control exists. Some technologists affirm their peers don’t consider the ethics when working on their passion. The concern with AI is so much so that top technology executives from around the world wrote an open letter to the UN “raising the alarm” to the concerns and potential threats AI can have if it’s used with weapons or for human targeting. Concerning? Yes.
So what do we need to know? Artificial Intelligence is around us, all the time. The implications are total optimization of our world. We have self driving cars, because computers can read the data, and program the cars to keep us safe. We have more customer service options, and faster shipping, and can find almost anything we want with an online search engine, because of AI. Still, there are positive ways to implement AI developments, and uncontrollable applications. It is our job to make sure our technologists, our futurists, our AI developers keep the ethics and interest of humankind at heart as they work towards full computer replication of human intelligence.