Can robots think and feel like humans do?
Or are they just programmed to mimic our responses to different situations?
Deep learning, machine learning. These couple of AI techniques were invented so that either virtual or animate machines can “self-learn” responses from analyzing inputted data.
By self-learn, it means that the robot will recall past information, identify patterns and connections from the given information, and then come up with a suitable response needed in a situation.
One common example of this is the recommendation list in a shopping app.
The AI integrated in the app will record and analyze an individual’s shopping purchases or history.
If the person usually buys apparel, then the AI in the app will recommend to that person other similar apparel that the person probably hasn’t seen before but might like.
It may also analyze similar shopping behaviors of a random set of individuals and recommend to them what else each other bought within the app.
It would seem like the app knows what you want, but apparently, it only aptly predicted the result based on your past shopping behavior.
It’s plain math. It hasn’t really become sentient or self-aware.
So far, robots made at the moment have already reached what is termed as “reinforcement learning.”
With this, some robots are learning the correct response through rewarding it positively when it has exhibited the proper response in a given situation.
It’s like teaching a dog new tricks.
When you train the dog to learn how to sit, you pat its head or give treats when he sits upon your command.
The dog will learn that hearing the command ‘sit’ and performing the action of sitting will provide him a good benefit.
Similarly, robots programmed with reinforcement learning in their chips would do the same.
If the robot exhibits the positive action in his environment, a human interpreter of that positive action will reward it.
The reward will be defined by the human interpreter, either by programming the robot that not hitting obstacles in an environment is good (1), while hitting obstacles is bad (0).
So the robot will learn to fulfill what gives it reward (1) rather than it earning zeros.
Reinforcement learning is applied to self-driving cars. But this technology is still considered as weak AI.
With strong AI, experts hypothesized that robots would achieve what is called sentience, or the robot having a mind of its own and is self-aware.
It would process information not only from a variety of experiences and its learnings, but will respond from feelings and beliefs as well.
Some have high hopes we humans can achieve such innovation, while for others, strong AI is just like speculative fiction, a topic of books and movies with imaginary settings.
We can only look forward to such inventions, but like the common phrase says: the possibilities are endless!
Heck, space voyage and touchscreens and the internet already materialized within a century!
And I would say, Wall-E, the robot capable of having feelings of love, must be cute in person.
What do you think about robots having thoughts and feelings?
Disclaimer: this article is for expositional and informational purposes only.
Thanks for reading! Have a great day. :-)