AI in the real world
While we haven’t got the likes of DARYL or C-3PO yet, computers with advanced AI are already in use in our everyday lives.
WHILE work on artificial intelligence hasn’t quite caught up with the movies yet, scientists are already making major advances in creating thinking computers, especially in recent years.
Here’s a list of the more notable real-world applications of AI.
Human vs AI
Probably one of the most well known AI computers of its time, IBM’s Deep Blue was programmed to play chess and was pitted against real-life opponents.
Deep Blue had many publicised chess matches against the then reigning world chess champion, Garry Kasparov and eventually beat him on May 11, 1997.
Following Deep Blue’s lead, Google launched AlphaGo, a computer that plays the board game Go.
The board game is said to be more difficult for a computer to win compared to chess because it has more possible positions than there are atoms in the universe, making it substantially more difficult for a computer to make a move by sorting through all possible moves.
What’s interesting is that AlphaGo was not specifically taught to play Go and learned to play the game using a generalpurpose algorithm where it picked up the rules by playing against humans and via simulations.
AlphaGo eventually beat Korean Go champion Lee Se-dol in a five game match, losing only once.
AI vs data
The latest incarnation of IBM’s AI is called Watson and it uses natural language processing and machine learning techniques.
What this means is that Watson can be given a load of data (which it can analyse faster than any human), then when queried by the user, it will extract the relevant data pertaining to the query and provide answers in seconds.
Watson is already being used in some hospitals in the United States where the AI software is used to sort through patients’ genetic data and pair them with available medical literature to find suitable treatments.
Closer to home, Hong Leong Bank Bhd actually uses Watson for the bank’s customer support by offering an online live chat which uses Watson’s AI to answer queries and make recommendations.
An example of Watson’s power is that it was used in the American trivia game show Jeopardy! in 2011, where it demonstrated how fast it could understand a question then provide the correct answers – Watson consistently beat the two human Jeopardy! champions in a three-day match.
AI vs machines
You’ve heard about those self-driving cars from Google and Tesla?
Well, self-driving cars would not be possible without a major amount of artificial intelligence.
All the companies working on self-driving cars today use different sensors to detect the road and obstacles, but it’s the AI that makes sense of all this information and controls the car to suit the situation.
In fact, Google is so confident of the artificial intelligence in its self-driving cars that the company has started building cars without steering wheels and pedals.
Self-driving cars are actually much closer to reality than you think – a company called NuTonomy is working with ride-hailing company Grab to run a trial for self-driving cars in a limited area in Singapore right now.
Of course, the technology isn’t yet perfect – a car from NuTonomy was recently involved in a minor accident (goo.gl/RpTc6w).
AI in your home
Possibly the most well-known forms of AI are of course, in your smartphone – Apple’s Siri, Google Assistant, Amazon’s Alexa and Microsoft’s Cortana are actually pretty sophisticated examples of artificial intelligence and machine learning that was not possible even a few years ago.
Just imagine – when you ask any of these virtual assistants a question it has to understand not just the words but also the context in which they are spoken and give you an answer in about a second or two.
So, when you say to Siri to “Create an appointment with my brother tomorrow at 1pm,” the AI has to first process what you are actually saying (taking into account your accent), then it has to understand who your brother is and then open the calendar to create an appointment at the appropriate time and date.
These virtual assistants also continuously learn about the user’s preferences and behaviour the more you use them.