AI will take over the world – and that’s a good thing

From Elon Musk to Stephen Hawking, there has been much discussion about machine intelligence. Far from taking over the world, the advances in AI are a necessary part of our evolution.

artificial intelligence / machine learning / virtual brain

The impending rise of intelligent machines and “Super AI” systems has been heralded in many headlines this year.  Many prominent business people and scientists have discussed the possibilities for what smart computers can do.  The conversations have touched on many areas, some mundane and others quite sensational.  In August of this year, over 100 prominent business leaders sent an open letter to the United Nations calling for a global ban on “Lethal Autonomous Weapons Systems”, which are AI-controlled instruments of war.  Even more recently, Stephen Hawking went on record by saying that “…AI could become the worst event in the history [of] civilization.”  Yet putting all of the hyperbole aside, no computer has yet to come forward with plans for world domination.

For people who do not closely follow the ongoing research and discussion around machine learning, it is difficult to understand why it might be relevant.  At the most basic level there are three fundamental differences between computers and humans that make machine learning and AI technologies important.  Understanding each difference can give us an excellent understanding of how machine learning works today and where the science might be headed.

Every computer system is gifted with perfect recall

Have you ever wondered why the Google search engine first became so popular or why it has stayed that way?  In general, humans struggle with the ability to remember information.  Although it wasn’t the first system to hit the Internet, Google made its engine such that it is easy, fast, and fun to find information on almost any topic.  The human brain is not designed to access all of its stored information when presented with an analytical problem.  In fact, while a brain can hold somewhere between 1 terabyte (TB) and 2.5 petabytes (PB) of information, the conscious mind can only process about 50-75 bits of information per second.  At that speed the human mind can only access a fraction of its knowledge when making decisions in real time.  A computer does not have those limitations.  Utilizing its silicon-based processors, a computer can literally access billions of pieces of information to apply against a real time decision-making scenario.  Everyone could see this advantage when IBM’s Big Blue supercomputer won its first world championship in chess.  But what about a business scenario?

If you’ve ever owned an American Express card you’ve likely seen the AMEX fraud prevention engine in action.  Through the use of advanced hardware and software, the company is able to compare each real-time transaction to every single transaction that a person has ever made.  They look at information like address, time of day, type of product, and many others to determine whether a purchase is normal or abnormal.  The AMEX AI continues to better understand each customer more intimately with every transaction that is made.  By using artificial intelligence and machine learning, American Express can harness all the available information every single time.  Comparatively, a human would only be able to consider a tiny slice of information and thus would be ineffectual at any type of real-time decision making.

It is often said that knowledge is power.  Machine learning has emerged as a very powerful way for computers to take over the burden of everyday analytics.

Computers never get emotional or make “bad” decisions

Several major companies including Google, Facebook, and Tesla are betting big on self-driving cars.  Automated cars represent a rare opportunity to vertically integrate many major components of an individual’s life – work, family, social, transportation – under a single brand.  The stakes are higher than most people know.  Intel corporation has predicted that the potential annual revenue generated by what they call the “Passenger Economy” will be worth $7 trillion by 2050.

Many people debate how automated cars and the roadways on which they drive could be constructed in such a way as to be safe.  However, the R&D experts know that by removing humans and their propensities for error from the driver’s seat, they can eliminate almost 93 percent of crashes.  The remaining problems become simple package and shipping logistics issues such as one might find in a FedEX or Amazon warehouse.

Artificial intelligence will be the “brain” within every automated car.  It isn’t a stretch to understand how a computer, with its ability to process massive data sets, would have more information than a human when making decisions on the road.  What makes the computer so important is that it will never get road rage, it won’t panic and jerk the wheel when an animal runs across the freeway, and it won’t get distracted.  What’s more, every single second of driving creates more data from which the AI can learn how to operate even better.

Unlike humans, computers are absolutely devoid of any type of emotion and that’s a good thing when you’re cruising down the road at 60 miles per hour in traffic.  An AI will drive “by the book” every single time without variance.

Computers may run out of power but they never get tired

Human decision making is greatly affected by biological concerns.  If a person is tired, hungry, or in pain their ability to exercise good judgment is impaired.  The U.S. Federal government limits by law the number of hours that a truck driver can work to just ten.  Airline pilots are also forbidden to fly more than 60 hours in any given week.  Statistically, the performance of a human in almost any activity starts to diminish drastically after a certain amount of time with no breaks in between.  Computers do not have that same limitation.  Artificial intelligence is already at work in many parts of our economy.  Since they do not have to rest, computer systems are becoming more important every year in running almost every facet of our lives.  Through machine learning practices the accuracy of AI systems improve on a daily basis.

One of the great fears about advanced AI systems is that they will eventually transcend humanity and rebel against their creators.  While that concept might forever be relegated to science fiction, the fact that computers can work tirelessly on our behalf, twenty-four hours per day, will make them incredibly useful to all of humanity.

There is much that we do not yet know about where the sciences of AI and machine learning will ultimately go.  But for just these three reasons this technology might just benefit us more than all of the science that has come beforehand.

This article is published as part of the IDG Contributor Network. Want to Join?

NEW! Download the Winter 2018 digital edition of CIO magazine