Artificial Intelligence has been gradually gaining traction in the field of computer automation systems. More and more consumer electronics devices have been implementing AI. From automatic vacuum robots to complicated software that the industries are using to replace real humans on the job – AI can have many implementations and implications.
Artificial Intelligence has various categories and types. The ones we see in science fiction movies, or even what we see in consumer electronics these days are just the a few of the various kinds of artificial intelligence that are theoretically possible.
Purely Reactive Intelligence
This is the most basic type of intelligence that is implemented in devices such as your automatic vacuum cleaner or even IBM’s Deep Blue – the AI which beat Kasparov at chess. Purely reactive AI looks at a situation and reacts accordingly. There’s nothing much complicated about it.
Your automatic vacuum cleaner has sensors that allow it to sense the presence of dirt on the floor and sensors to see where the floor ends. So, when your vacuum cleaner robot senses dirt on the floor, it simply proceeds in that direction to clean the floor.
Deep blue is a more advanced version of your robotic vacuum cleaner since it also takes into account all the possible outcomes that can arise out of one of its decisions. When it’s Deep Blue’s turn to make a move, it analyzes the chess board, looks at all it’s possible moves, checks out what moves it opponents can give for each of its moves. That’s a large number of computations and the with each level deeper, the number of computations needed grows exponentially.
One thing that a robotic vacuum cleaner and Deep Blue have in common is that they do not need to learn the task they are assigned to do – that is already programmed. There is no learning involved. Like Deep Blue, Google’s Alpha Go that plays the game of Go is a reactive intelligence system. So it your smart refrigerator, your smart thermostat and smart home lighting system.
Limited Memory Systems
This s the next step up the ladder. Here, the AI has a limited amount of memory that it can use to learn things. Now, this limited memory can be even terabytes of storage. This storage contains programmed representation of the world that is perceived by the intelligence. The more the intelligent system looks at this data, the better it can do the task that it is assigned to do.
The best example of this scenario is a self-driving car. The intelligent system that powers a self-driving car needs to learn how to react to a situation on the road in real time. For example, if there’s a pothole on the road, the AI will initially do nothing that drive over that, only to experience an unwanted jerk. The AI will record that, and the next time it sees a pot hole, it will try go slower. Over time, based on the location and size of the pothole it sees, it may drive the car around.
Personal digital assistant, like Google’s Assistant, Amazon’s Alexa, Microsoft’s Cortana or Apple’s Siri have a similar AI implementation. That is why they get better at listening to your commands over time. This is also why all these companies collect huge amounts of user data, raising privacy concerns.
Theory of Mind
Another step up the ladder and we are looking at intelligent systems that can understand human thoughts and emotions. This type of AI can comprehend human thoughts, motives, intentions, expectations and can interact socially. We have seen such sort of AI in movies only. There’s C-3PO and R2-D2 from Star Wars. Also, there’s Sonny from the movie I, Robot.
However, there are machines that have been made to understand human emotions. Facebook has been developing an AI that will help prevent suicides. This is far from what we seen in movies, but, with the help from other humans, this AI can spot a troubled user and contact his or her friends to reach out to him or her. There are also other systems that are being designed to human emotions, but they are still research work.
Super intelligent, sentient, conscious – this is the AI of the future. These are machines that contemplate their own existence, they are aware of their internal states and they understand feelings of others. Eva in the movie Ex Machina released in 2015 is such an example.