Artificial Intelligence and Major Challenges | AI Challenges

What is AI or Artificial Intelligence?

Artificial intelligence, this term has been getting thrown around in the media a lot these days. But what exactly is it? The AI or Artificial Intelligence is a branch of computer science wherein essentially a machine is able to make human level decisions; more specifically AI is used typically in conjunction with machine learning to perform various tasks like pattern recognition, visual interpretation, speech recognition etc.

In essence though, AI is like like a model of the human brain. Just like the human brain AI uses neurons and learns based on experience. There’s a lot of other branches of AI like neural networks and statistical learning but all this end up there. To keep it simple there’s no doubt this technology makes our lives easier.

AI In our Life

Virtual assistants like Google Assistant, Alexa and Siri are some obvious examples. Some of which you may not have realized, come in the form of your online advertisements music. Even Netflix movie recommendations which are all produced through an AI learning algorithm based on what it discovers through your hobbies interests and search history.

Impacts of AI

There are many issues that will arise the more widespread artificial intelligence gets. Contrary to popular belief, the most concerning issue regarding the development of AI is not the robots that are gonna take over the world. That is very unlikely to occur within this lifetime.

Instead, the most concerning issue will be the loss of jobs. As AI gets more advanced human labor is no longer necessary. For many tasks in the workplace a recent report predicts that by 2030, as many as 800 million and 30% of all jobs could be lost worldwide due to automation.

A few of the most threatened industries include warehouse jobs, manufacturing jobs and delivery services which we can already see happening today.

There are songs that are developed solely using AI. Daddy’s Car was the first ever song composed using purely artificial intelligence.

Daddy’s Car: A Song Composed by Artificial Intelligence

Using flow machines developed by scientists in sony’s CSL lab, the artificial intelligence technology was able to piece of the song based on the style of the Beatles, following the style by viewing it as a computational object. In order to do this, the flow machines had to imitate and combine the works of many influential creators until a unique style was formed.

Songs like this cost controversy regarding whether or not artificial intelligence was even threatening artistic industries. Others argue that the development of artificial intelligence can potentially become unethical, because in addition to the major job loss, it is unclear how to distribute the wealth acquired from these machines and there’s a lot of uncertainty regarding its safety and consequences. If one person receives all the profits made this will create an imbalance economy creating more and more monopolies.

Despite all this is not all going to be bad. New jobs will be created existing roles will be redefined and workers will have the opportunity to switch careers. Reports show the worst effects on this transition can be mitigated if governments begin to take an active role.

In order to fix these issues, we need to raise awareness to plan ahead and most importantly get the government involved. This issue is going to continue to grow as the development of artificial intelligence grows exponentially. We need to make sure there are still enough jobs available and make sure an AI doesn’t negatively interfere with too many industries . We can make this happen by petitioning and adding regulation laws to ensure this technology doesn’t end up backfiring.

AI Challenges

Two areas that AI isn’t very good at are Improvising and Contextualizing, both of which come very naturally for humans. Everyone, of every skill level, can Improvise in a way far beyond any machine.

Consider an example. What If the door handle breaks off in your hand? What will you do? Well, I dont think you would simply stand there dumbstruck. Do you?

No, you will try to figure out a way to get the door open. What if you are building a robot plumber?

Every house is different, and there are so many variants of equipments and products, and there are almost limitless things that can go wrong with your plumbing work.

Image result for funny flow chart

A human plumber doesn’t have to train on every variant of every equipment and models. So that when the owner of the building calls our robotic plumber and say “We need to modify the bathroom upstairs by making it usable for handicapped people, with very little change as possible” the plumber doesn’t panic and think, “I cant! I haven’t trained on that model yet.”

Image result for funny ai

What about Contextualizing? Well, here also AI also has a hard time.

Imagine you are in a movie theatre, a sad scene is going on, the girl next you is crying and her boyfriend is comforting her, you wouldn’t have any trouble making a sense out of that scene.

But what about a computer? A computer capturing and processing this same scene. For a computer, these are all vectors. Vectors with endless possibilities. What they are seeing is simply a collection of ones and zeroes. In order to make sense out of it, it needs more data; more and more ones and zeroes to reach a conclusion. Do you see the problem here? Just imagine how easy is for us to figure out what a photo says.

Check out the below pictures

Image result for pictures
Image result for romance
Image result for fun
Image result for fun

By looking at it just for a second, you knew what these pictures meant We can look and say: “Oh, they are skydiving”. “Ooh, so sweet. Nice Couple.” “Having fun time with friends.” and so forth. These were all easy for you, right? Why? Because we have the cultural context to decipher it, which AI lacks. These are just a couple of the many challenges that AI researchers struggle with today.

Written by Jithin


Leave a Reply

Your email address will not be published. Required fields are marked *





Arduino Game Board

Adafruit PyGamer | Circuit Python, Arduino & Makecode Arcade