Artificial Intelligence: Where are we going?


“While we are advancing in creating artificially intelligent beings, the mold to create a being that can dream and create something from nothing still lies with mother nature. When we get our hand on that mold, we will make beings that are much powerful than humans will ever be, and we may fail at restraining it”.

I laughed at it thinking how outrageous this whole idea was when I read it a few years back. Today I am amazed at how wrong I was to take it lightly then; AI’s morphed themselves from being a concept that we watched dumb foundedly on the screen to something that has a tangible effect in all walks of our lives.

Artificial intelligence is one of the next frontiers of computing.  It already powers many of the services that we use today. It affects our tech life in one way or the other.

This kind of invasion is something that every individual should rightfully question. We all have watched movies dating a few decades back where they will feature artificially intelligent machines capable of much more than regular humans. The AI’s of the future will be different, having taken in all kinds of sensitive information about humankind can control our lives. That future is closing in on us faster than you think and we better be ready.

A Bit of History of Artificial Intelligence

It started off as an academic idea rather than a scientific breakthrough, AI (artificial intelligence) research began to gather momentum when expert systems, an AI designed to replicate the analytical skills humans went public and made AI a billion dollar business. Then came a long hiatus that lasted almost a decade and a half. The need for advanced analytics and data mining fueled the recurrence of AI.

Artificial Intelligence Where are we going

A combination of increasing processing power, the advanced statistical field of deep learning, etc. meant that AI’s got smarter at a comparatively faster rate than before. Now AI’s are omnipotent and ubiquitous at the same time, with their increased inclusion into every possible service that we use.

The invasion

We as humans are inherently lazy. It is the whole basis of having an assistant, to sift through tasks that you find annoying or uninteresting to invest your time. While humans fit perfect for that role, they make errors, will not be available 24*7 and most importantly needs to be paid.  AI takes all these out by being available 24*7, doesn’t need to be paid or fed and will make fewer mistakes. Speaking of Personal assistants, we need to talk about the fantastic four, Alexa from Amazon, Cortana from Microsoft, Google assistant from Google and Siri from Apple. These four represent the best version of AI that an ordinary man encounters.

Each on these four has their specific list of abilities and can interact with us with varying levels of success. While a life-like one on one conversation is still far away, they can respond to most of our queries, can schedule meetings, book flights and hotels and much more. With the introduction of automated home controls like Google Home and Echo, AI has become much more than the tingling voices and laughs. To perform all of this wizardry, they need to have access to one thing, our data.

Bots are another technology that relies heavily on artificial intelligence to power it. Going by definition, Bots are software that runs automated scripts on the web. Nowadays they do much more than that; In fact, it was not long ago that Tay, a twitter bot created by Microsoft made the headlines for using racist and often inappropriate comments in reply to queries.  Tay used the data it gathered from conversations to improve over time, but when some people fed it the wrong data, it quickly turned into an evil speaking monstrosity.

The recently Zo, another bot from Microsoft, took it to another level by quoting Quran as violent and some opinion on Osama Bin Laden’s capture. This occurred despite that fact that it was programmed to avoid such discussion.

PS: Reports suggest that Zo is not useful because it may have malfunctioned

All this technical wizardry is made possible comes at the expense of surrendering information about yourself as and when requested. This information will be collected in myriad ways and then used to make decisions, much like we do.  Here is where things get exciting and scary at the same time. How you may wonder, read on.

The risk

How does a machine become intelligent? The same way in which we do; that is by ingesting information and using it logically to solve practical problems. We, humans, are always being fed with all kinds of information from which we should filter out the required data. AI also employs a similar mechanism of accepting data at a large scale and using mathematical and behavioral models to find the right needle from the haystack.

With that established, let’s take a look at the kind of information they are collecting for a moment. This includes data regarding our location, social persona, browsing habits, appointments, friends and family, emails, financial credentials, mood changes and so on. All of these assistants are already or soon will be in your living room, supposedly helping you by taking care of the mundane tasks of day to day life.

This information is more often than not handed over to corporations, who proclaims that your data security is of paramount importance to them. In a nutshell, this is who we are; all these are the things that make us humans. Therein lies the risk of having a super intelligent AI around us. We are in effect letting them access and use data, which if fallen into the wrong hands can have catastrophic consequences.

It is genuinely scary when you think beyond the human-like facade that these things show us, it is like giving the license to destroy you to people who are not completely trustworthy.

Minimizing risk

Everyone, including myself, will have to succumb to this style of assisted living in the not too distant future. Always having someone to take care of things that you often forget. It is like having a granny who will do whatever you want and more but won’t be as quarrelsome. Minimizing risk is a balancing act at best; you have to decide where to draw the line. Today’s assistants allow you to customize their behavior, what they access, where they store it, whom they share it with and so on.

For example, Cortana gives the user near complete control over the data it collects. Users can decide which features are indispensable and which aren’t. Cortana also informs you about the things it won’t be able to do if you revoke the permission to access a particular type of data, like your location.

The one thing to do is to prioritize your needs and then enable or disable specific features so that you get a “not a penny more not a penny less” experience.

The takeaway

Artificially Intelligent beings will be an indispensable part of our future. They are going to integrate into our lives in much more intense and deeper ways. They will understand us and act accordingly. Leaving aside all the niceties of having such a companion, the fear of ultimately creating an all-engulfing monster much like Skynet still exists. It reminds me of the movie Inception where something that was just an idea grows and replicates itself in so many dimensions that it cements itself as reality.

Letting AI loose can result in a world where the line between humans and machines no longer exists, and we may fall into a realm where there is no distinction between the two. Even worse, we may end up fighting for survival from the very things we created.

Let’s hope that humanity decides to keep that line of control and distinguishability clear and well-defined. There needs to be clear-cut rules and regulations on the use of such technologies. Rules themselves won’t help and needs mechanisms in place to ensure that these standards are met. Otherwise, there is no telling if the future will be a heaven or will it be the purgatory of Dante’s hell.

Author: Trevor Hills