What are the issues associated with robot development using artificial intelligence and what are the advantages of artificial intelligence?
Robots are nothing more than computers programmed to look like they have an autonomic nervous system. Since they are programmed to perform specific tasks, they can be made to do anything (Logan & Braga, 2020). The most basic robots are programmed to obey simple commands, like lifting a box and putting it back down. This is sufficient for a lot of activities, such as handling industrial products, or security.
The
Issues Found In Developing Robots With AI
But
there are many tasks that robots cannot yet do. For example, they are
not good at interacting with people. They will sometimes bump into
people and scatter objects. They don't understand jokes, or small
talk, or facial expressions (Morgan,
2017).
The
basic problem AI programmers face is that computers are not very good
at dealing with uncertainty, and robots are no exception.
There
are several reasons for this. The first, and most obvious, is that
computers don't understand the world the way we do. We don't say "X
is a teapot," we say "X is a teapot." We say "X
is of shape Y," not "X is a teapot-shaped thing." We
say "I know that Y is real," not "I know that Y is the
case."
Artificial intelligence is about more than robots.
It's about software. If software can do something, then a computer
can do it. If software can make something happen, then a robot can do
it.
In software, artificial intelligence is often called machine
learning. A program that uses machine learning is called an expert
system.
Machine learning is exciting. But there's a problem: it
requires that we first teach the computer how to do a task.
Teaching
a computer to do something is extremely difficult. It requires that
we be able to explain (or code) task knowledge in computer language.
If we can do that, then a computer can learn.
We can teach a
computer to recognize words, for example. But a human can do it more
easily. Humans can learn one word at a time. The computer has to
learn everything at once.
The problem gets worse. Computers
can't make mistakes. Humans make mistakes.
Advantages
Of Artificial Intelligence
They can do work that costs human labor. They can do work that is too
dangerous for people to do.
But robots can do things that people
can do better. They can work longer hours. They can work at night
when people are not. They can work when there are not enough people
to do the job. They can work where it is not pleasant to work. The
things we now do with robots, we might one day do with humans. But
doing them with robots is a lot cheaper, and that will make them more
valuable.
But robots can do things people cannot do. They
can learn. They can solve problems following their programming. They
can adapt to new situations. They can work in places where people
prefer not to work. a lot cheaper, and that will make them more
valuable.
Computers are great at calculating formulae, but
they can't build anything. They can't decide for themselves what to
build and how to do it. And they can't make predictions.
Artificial
intelligence is about making computers that can do both of those
things.
But in general, artificial intelligence isn't about
"making" anything. It's about helping people do what they
already try to do.
So the first thing you should do before
thinking about building a robot is to think about what it can do, and
then try to build that.
References
Logan, R. K., & Braga, A. (2020). AI and the Singularity: A Fallacy or a Great Opportunity? Mdpi AG.
Morgan, B. (2017, August 17). 10 Things Robots Can’t Do Better Than Humans. Forbes. https://www.forbes.com/sites/blakemorgan/2017/08/16/10-things-robots-cant-do-better-than-humans/?sh=54103fe8c83d
No comments:
Post a Comment