What Exactly is Artificial Intelligence

HAL : “I’m sorry Dave, I’m afraid I can’t do that.”

Mother : “The ship will automatically destruct in T-minus five minutes.”

The realm of science fiction (2001: A Space Odyssey, Alien) is littered with examples of artificial intelligence and/or artificial general intelligence. Today we live in a world where these sentient machines are nearly within reach and each day brings about breathtaking advancements.

The 1956 Dartmouth Summer Research Project on Artificial Intelligence is considered by many to be the first formal gathering of professionals to discuss what would become the field of Artificial Intelligence (AI). The term itself is credited to John McCarthy, a computer scientist who organized the event. The Merriam-Webster dictionary defines AI as: (1) A branch of computer science dealing with the simulation of intelligent behavior in computers. (2) The capability of a machine to imitate intelligent human behavior.

AI is comprised of sub-fields / semi-correlated disciplines such as Artificial Neural Networks, Computer Vision, Expert Systems, Robotics, Speech Processing, Machine Learning, Natural Language Processing, Evolutionary Computation and others depending upon who you ask. The examples of AI are all around us – Apple’s Siri, Amazon’s Alexa, Microsoft’s Cortana, IBM’s Watson, etc., to name just a few. Google, Facebook, PayPal, UBER, Netflix and countless other companies are grabbing onto the technology with both hands – and in many cases helping to define it. The use cases of AI are almost beyond comprehension such as: estimating obesity rates based upon urban neighborhood satellite imagery, determining personality traits from eye movements, predicting the likelihood of having heart disease more accurately than physicians, identifying potential criminals using facial features alone, etc.

AI will touch virtually every aspect of our lives and likely contribute trillions to the global economy within the next decade. Thought leaders are comparing the impact of AI as that equal to the steam engine and even electricity. Amazingly, the data will only get more voluminous, the hardware and processors faster, the algorithms more precise, and the impact more significant. The technology will be present in every industry and both displace and create a massive number of jobs. The race for AI dominance is not surprisingly at the nation-state level whereby bold predictions and assertions are being made by former cold war adversaries. AI will advance the human race in ways that we cannot even comprehend today. The cures for diseases will be unearthed, crime detection and prevention could look something like the film “The Minority Report”, and you may have a humanoid robot in your home.

AI as we know it today is largely based upon machine learning and considered “Weak AI” versus “Strong AI” whereby we are still very far from emulating human intelligence. The ultimate end-game for many AI practitioners is “Strong AI”, Artificial General Intelligence (AGI) and the “Singularity”, all of which will be discussed in later articles. The ability to achieve AGI is still being debated by many and its realization is heralded by some as the beginning of our demise as a species. I like to tell people that we might be able to embed our values and morals into AI (Weak AI) but AGI (Strong AI) will essentially be able to rewrite itself and not every human shares the same moral code base – pun intended. Countries are already talking about developing unmanned AI submarines and other weapons of war, let alone criminals and unfriendly actors who could use the technology for their own purposes. The sobering fact is that anything we create can be used against us whether directly or indirectly.

AI has been pursued for many decades thanks to brilliant and persistent professionals such as Geoffrey Hinton, Andrew Ng, Michael Jordan (not that Michael Jordan), Fei-Fei Li, Yoshua Bengio, Yann LeCunn and many others. The most common question seems to be: why are we hearing so much about AI and machine learning now? The primary reasons are over the past few years we have been granted access to enormous amounts of data, faster hardware and processors, and in some cases, recently refined or “solved” mathematical, statistical algorithms. The job market for AI and machine learning professionals is astounding, as is the very much related data science market. Please recall that much of AI as it exists today leverages machine learning which typically requires massive amounts of data in order to be trained. Indeed, freshly minted advanced degree holders are commanding $200K to $300K salaries, professors are plucked from top universities and treated like rock stars, and data scientists are thought to hold the sexiest jobs around. I can tell you that if I were headed off to college today my focus would be a bachelor’s in computer science / artificial intelligence, minor in data science and mathematics, all the way through to PhD.

Artificial intelligence cannot be stopped but it can be regulated, and many are already calling for legal framework discussions. For example, to prohibit discrimination, autonomous intelligent weapon systems, the humanizing of robots, etc. In addition, weighty issues such as who would be legally responsible if the intelligence proves to be incorrect in a medical diagnosis and/or drug prescription, when the self-driving car causes an accident (sadly it already has resulted in a fatality), and thousands of others. I believe AI will be seen as the most impactful and beneficial advancement in our species within the next two decades. I also believe that we need to be extremely careful regarding how it gets used and how it advances. After all, we can only control what we ourselves do – and we often struggle with that alone.