Artificial Intelligence or AI for Dummies

Chances are that you have used Google Maps to get directions. It is almost perfect and however obscure your destination it gets you there.

Before Google Maps we would remember the route and a few landmarks but now most of the time we just feed in the destination, relax, and go back to whatever we were scrolling. Lo & behold, Google lady directs the driver and ensures we reach exactly wherever we wish to go.

Some of us use Alexa or Siri voice commands. She is an obedient assistant, does as she is instructed without fail, and has no irritability no matter how many tasks you give her.

From time to time on your phone, you get a memory of a bunch of photos of yours or your family. That is AI recognizing a pattern of your face and grouping them all together.

If we have used these Apps even if inadvertently we have used AI or Artificial Intelligence.

Artificial Intelligence is trying to make computer systems act intelligently like a human brain. It tries to make a computer mimic a human being with reasoning power, solving problems, and making decisions.

How Does This Happen?

The question is how does a machine learn? It is given a huge amount of data and asks it to sift through it. The computer program is “TRAINED” by humans by giving labels and a set of instructions.

The program then looks for patterns and categorises them for images and sound. Humans supervise the learning and correct the program when it comes up with wrong results.

For eg the machine is fed thousands of pictures of a car, it understands how it looks and can identify from a fresh set of images which ones are cars. Sometimes a truck may look like a car and if it makes a mistake, it is nudged and over a period of time just like a human the model understands what to look for.

The machines are now so smart you can ask them to produce a picture of a flying car with a 1950 Mumbai backdrop painted in the style of artist MF Hussain or Picasso. The creativity and accuracy of the outcome is quite extraordinary.

Not bad, right?

What Are ChatBots or Large Language Models?

Most of us have dealt with customer service bots, or chat boxes that give us options and process orders on merchant sites. Some chatbots are scripted and others are more complex and can provide reliable contextual responses.

These Large language models are computer programs that are fed with humongous volumes of text. Just like we learn a language, the computer understands the words and forms the response in the correct context without human intervention.

They are not 100% perfect yet.. but pretty much up there.

“In 10 years, I think we will have chatbots that work as an expert in any domain you’d like. So you will be able to ask an expert doctor, an expert teacher, an expert lawyer whatever you need and have those systems go accomplish things for you.”

Sam Altman – CEO of OpenAI, creators of ChatGPT

These are huge machines that crunch massive amounts of data patterns and identify the most likely answer if a question is posed to it.  These programs can do amazing feats like they can predict the weather, analyse hundreds of scans and spot cancers that doctors can miss, compose poetry, write songs, and unravel games like chess. You name it, train it, it will deliver, all this in seconds.

The more complex AI, which is called Artificial General Intelligence, works much more like a human brain. Just as the brain has billions of neural pathways makes connections and is capable of learning things almost independently. They use a form of machine learning called deep learning that uses ANNs or artificial Neural Networks or SNNs, simulated neural networks to develop algorithms to find solutions.

Just like a human brain it is capable of rational responses and can self-learn. Scientists who have created these programs are not very sure how they work or what are the limits to AI’s learning ability.

The ethical question is that these machines become super smart and it may not be possible to control them. The fear of a dystopian reality out of a sci-fi movie where robots go rogue and destroy our planet is not farfetched.

What is Chatgpt?

If any of you have used Chatgpt, you will know it’s fascinating. When you google something, it comes up with search results. You click on the results and draw your conclusions. If you ask ChatGpt it just gives you the answer. Albeit sometimes they are plain wrong ones.

ChatGpt was created by a company called OpenAI now owned by Microsoft, Google has its own Gemini & Meta LLAma3 are all LLMs or large language models.

All of them are imperfect as yet but are making rapid progress. Some of the inherent human biases and prejudices have crept in which need to be cleaned and tested over some time.

How Does Self-Driving Work?

Just like a dragonfly, with 360-degree vision & sensors on its wings to help it maneuver, the car is equipped with AI models with sensors. The model is fed with thousands of hours of training and it can make decisions in the real world while driving. The AI can sense if there is a bicycle nearby or a pedestrian and like a good driver avoid collisions. It is uncanny.

Autonomous cars are almost a reality. In San Francisco, there are paying customers renting self-driving cars. Scary for the future careers of drivers if it is proven to be as good or as bad as a human driver.

Is Artificial Intelligence Safe?

We don’t know. It will have to be made safe by laws and self-regulation.  These machines mimic human behavior, so we have to be like good parents setting good role models so that it is nudged to behave correctly. It’s a funny analogy, treating a machine like a child.

Warren Buffett compares AI to a nuclear weapon and says it can be used to scam people. Might as well be true.

Whether we like it or not, AI is here to stay, we will have to adapt to it. In the near future, it may result in significant job losses as many repetitive, boring jobs can be done far more efficiently with AI.

Just like how we learned about the internet, we will understand AI too & become more productive. That is if it hasn’t taken our job from us.

I could have written this blog or just asked  Chatgpt to write a short blog on AI.

Who knows maybe this one is written by it. Can you guess? 🙂