Artificial intelligence (AI) is a repeatedly misused term.
Coined in 1956 by John McCarthy, it was originally intended to define an independent machine agent that can take actions to maximize success toward a particular goal, with human-like functions such as learning and problem solving. AI can be broadly categorized as ANI (artificial narrow intelligence), AGI (artificial general intelligence) and ASI (artificial superintelligence). Almost all of the AI systems we see today align under ANI — e.g., IBM Watson, Deep Blue, a calculator, even the device you’re reading this from all fall into that category. All are built to perform specific functions, but are not quite at a human level.
Mind vs. machine
Watson could easily find Jeopardy answers, Deep Blue managed to win a chess game against the reigning world champ (recent bug stories aside), a calculator can quickly solve complex mathematical problems, and the device you’re using to read this can perform specific tasks at a high processing speed and with just one plausible output — yet certainly this is not how a human brain functions. Machines today may be able to produce faster and more accurate results, but ultimately, they cannot think on their own.
Humans have a wider range of capabilities — we may favor Fight Club over Batman vs. Superman: Dawn of Justice (or vice versa), we can argue over Mona Lisa’s expression, and we want to order extra french fries for no logical reason. Simply put, we’re complex. It’s in our DNA to have emotions, express feelings, understand subtleties in expressions and gestures, have irrational fears, crave certain foods, and hold countless other qualities that binary numbers will take a while to decode.
To learn human patterns is a colossal task, though this could eventually be achievable. As humans, we still haven’t mastered it; just look at the wars and social injustices that have plagued humanity for centuries.
Mankind adapts to technology
With any technological advancement that has hit civilization — take the Industrial Revolution, for instance — there has been a natural fear of job security. Many jobs were lost. Even so, we still adapted. We created more jobs. We got better at a lot of things. We increased our efficiency in many areas. We optimized energy consumption.
Around 30 years ago, when computers were first becoming part of our work lives, our work environments changed dramatically. Humans could now automate many daily work activities. Banks, factories, government entities, medical facilities, and more used computers to do things like increase data storage, facilitate communication, and track bookkeeping records. This allowed humans to forgo roles with redundant tasks and focus on where they could add value with qualities that only a human could have. Human focus areas then shifted toward improving technology, security, creativity, efficiency, and other areas where our ability to see beyond 0 or 1 could be leveraged.
Job security and the rise of the machines
As computer usage increased, there was another wave of fear among people for their job security — similar to what we saw during the Industrial Revolution. The “job security” phrase was at its peak between 1980 and 2000, according to Google’s Ngram Viewer.
While computers and technology have a stigma attached in terms of job security, there’s an obvious flipside: the improvements computers have made to our quality of life. It creates somewhat of a Catch-22 in that we want the efficiency technology allows and want it to keep advancing, but not to the point where it could do an alarming number of our jobs without us. Historically, computers and technology displaced certain types of work, but they created many new jobs.
Take search engine optimization (SEO), for example. Sure, technology can now highlight your relevance to a certain topic on the web, but it takes someone who understands the underlying logic, and this paved the way for SEO experts who have some of the most sought-after skill sets in the business world today. And there are other roles where we’re still not ready to bow out of the field, despite technological advancements: Though a plane today can run mostly on its own, we wouldn’t dare step into a one without a human pilot physically in the cockpit.
As the AI layer reaches a human level of intelligence (artificial general intelligence, or AGI) in terms of computational power and ability to think, job security will get a lot of attention again. When AI becomes smarter than all humans combined (artificial superintelligence, or ASI), there are multiple scenarios it could introduce. For one, it might be the time where we can all kick back, relax, and let the machines handle our daily grind while we focus on areas of our choosing.
For instance, complex decision making in government, education, medicine, and other fields will still need human assistance before final decisions can be made, at least for the foreseeable future, even if AI is used to execute on most of the grunt work. Alternatively, the paperclip maximizer scenario could prevail, where machines could be programmed to perform seemingly innocuous tasks without an ethics layer, posing an existential risk.
Some computers have already matched human calculations per second, if not more. The biggest hurdle with such computers is the massive area they occupy and their need for a high level of power sources to function. In Ray Kurzweil’s The Singularity Is Near, he explains that by 2045, $1,000 will get you 1,000,000,000 (yes — one billion) times the power of all human brains today, combined. The year 2045 suddenly doesn’t seem so far away!
Look at the difference between us and many other animals for perspective on what this could look like: dogs and their ability to sniff out subtle smells, parrots recognizing colors, rats managing to find the only food left in the house. These animals can identify objects and react to threats, but can never self-reflect. The intelligence gap between these animals is fairly small, and even then their capabilities vary significantly. A chimpanzee, which shares 99 percent of the same DNA with us, is incapable of understanding that a noisy object that flies over them with stiff reflective wings is a manmade object and is not a part of nature, like they assume. That 1 percent counts for a lot. Imagine the scale of things that a super intelligent layer can accomplish — it will be much more advanced than human DNA. Just as we don’t communicate with ants or snails, it will be interesting to watch AI unfold and look down on us. Their thoughts will be entirely above our heads.
AI in 2016
But until then, what does AI look like today? We have Google Assistant, a decent helper. It can help us reserve a table at a restaurant, book tickets for a show, set reminders, send calendar invites and emails, check schedules, help us run errands, and prioritize our work. Can it replace an actual human assistant? Maybe not. But it may be able to do a few things better and faster.
Google Assistant is by far the most advanced AI layer compared to Cortana and Siri. They are all programmed to learn human patterns such as our choice of apps, who we are most likely to call in our contacts, and what our general preferences are. They can even understand what we ask them to do — sometimes.
Even if they don’t understand, they know they can substitute an answer with a search engine. That’s where Google has an upper hand. Its new premium smartphone (the Google Pixel) has taken a giant leap toward improved AI. Facebook is also reportedly working on an AI project, and so is Samsung. With their heaps of data on each of our interests and social circles, it will be interesting to see how they fare compared to our existing assistants today.
The infinite capabilities of AI
The potential of artificial intelligence has no ceiling; machines can be programmed to make themselves smarter and make coding changes to themselves at exponential levels. Stephen Hawking and Elon Musk have warned us of its dangers, and of the fine line that we are walking. The scale of what AI can do is unimaginable — if it doesn’t bring about our doom, it could help us find a faster solution for climate change or dramatically increase our lifespan. The potential of AI hits both extremes, so we have to tread these new waters with caution.
AI will inevitably replace a lot of jobs, sometimes putting entire fields out of work. But we will adapt — just like we always have. We adapted when the Industrial Revolution hit us. We will find new types of work by taking advantage of human capabilities. Highly skilled workers will still be needed. Menial tasks will increasingly be handled by AI, which will bring new opportunities that only a human can fill. As the world moves faster and as we build new technologies, we’ll gear up and learn these new skill sets.
Our choices control our destiny as we test the limits of what we can create. It’s the dawn of AI, and it’s going to be fun.