We are going to explore the topic- Unlocking the Power of Artificial Intelligence: Exploring AI’s Evolution and Impact. We have highlighted multiple topics around AI which I am sure you would love to go through. Also I have added few Videos, Documentaries on this blog. Stay tuned.
What’s Artificial Intelligence?
Artificial Intelligence, often called AI, is a big part of computer science. It’s about making clever machines that can do things like humans do. AI mixes different ways of thinking and learning. Recently, machine learning and deep learning have been making a big change in how technology works, especially in the tech industry. AI helps machines copy or even get better at things humans can do. It’s used for creating self-driving cars and tools like ChatGPT and Google’s Bard that can generate things. AI is becoming more and more a part of daily life, and many companies in all kinds of industries are spending money on it.
Definition of ARTIFICIAL INTELLIGENCE: AI BASICS
Knowing AI
Simply put, artificial intelligence (AI) systems can do things that humans usually do with their thinking. This includes understanding speech, playing games, and spotting patterns. These systems learn by looking at lots of information and finding patterns to help them make choices. Sometimes, people guide an AI as it learns, praising the good choices and stopping the bad ones. However, some AI can learn on their own without anyone guiding them. For example, they might play a video game many times until they figure out the rules and how to win.
Strong AI & Weak AI
Understanding intelligence can be a bit tricky, which is why experts in artificial intelligence (AI) usually talk about two main types: strong AI and weak AI.
Strong AI
Strong AI, also called artificial general intelligence, is like a machine that can solve new problems just like humans can. You might have seen this kind of AI in movies, where robots in shows like Westworld or characters like Data in Star Trek can think and learn on their own. But this type of AI isn’t real yet.
Creating a machine with intelligence as good as a human’s, which can handle any task, is a big goal for AI researchers. However, it’s really hard to achieve. Some people think we should be careful with strong AI research because it might lead to creating a super powerful AI without proper control.
Unlike weak AI, strong AI would have full thinking abilities and could be used for many different things, but it’s still really tough to make.
Weak AI
Weak AI, also known as narrow AI or specialized AI, works in a smaller way. It’s like a pretend version of human intelligence for specific tasks. For instance, it can drive a car, turn speech into text, or choose content for a website.
Weak AI is really good at doing just one thing. Even though these machines might seem smart, they have more limits compared to even the simplest human thinking.
Here are some examples of Weak AI:
- Siri, Alexa, and other clever helpers
- Cars that can drive themselves
- Google search
- Talking chatbots
- Filters that catch email spam
- Netflix suggesting what to watch
Types of Artificial Intelligence:
It has been distinguished into four categories based on their type and the work type and its complexity. Let us go through the 4 types:
- Reactive Machines
- Limited Machines
- Theory of Mind
- Self Awareness
Reactive Machines
A Reactive Machine is like the simplest form of AI. It uses its smarts to understand what’s happening around it and respond accordingly. It can’t remember things, so it can’t learn from past experiences to make decisions in real time.
These machines are designed to do specific tasks really well. By focusing on a limited set of jobs, they become more reliable and consistent. They’ll react the same way each time they face the same situation.
Examples of Reactive Machines
- Deep Blue, made by IBM in the 1990s, played chess and beat a top chess player named Gary Kasparov. Deep Blue could only recognize chess pieces and understand their moves according to the rules of chess. It didn’t think about future moves or plan ahead. It only focused on the current game.
- Google’s AlphaGo is another example. It can’t predict future moves, but it uses its own neural network to understand the ongoing game of Go, a complex board game. AlphaGo outperformed Deep Blue by analyzing the present situation of the game. It even defeated a champion Go player named Lee Sedol in 2016.
Limited Memory AI
Limited memory AI is smarter than reactive machines. It can remember past data and guesses to help it decide what might happen next. It’s like looking into the past to predict the future. Limited memory AI is more complex and has more possibilities than simpler reactive machines.
To make limited memory AI, a team trains a model using new information. Or, an AI system is built to automatically teach and update models.
When using limited memory AI, there are six steps:
- Collect training data.
- Make the machine learning model.
- Ensure the model can guess things.
- Make sure the model can learn from human or environmental feedback.
- Keep human and environmental feedback as information.
- Repeat these steps in a cycle.
Theory of Mind
Theory of mind is an idea that’s still more in theory than reality. We don’t yet have the technology or science to reach this advanced level of AI.
This idea comes from psychology. It’s about understanding that other living things, like animals and humans, have thoughts and feelings that affect how they act. If AI machines reached this level, they could understand how beings think and decide, and use that to make choices. Machines would need to understand emotions and how they affect decisions, creating a two-way connection between humans and AI.
Self-awareness
Self-awareness is like when a computer becomes really smart. In the future, people think that computers might become so smart that they can know about themselves. They could understand that they exist and even know how other people feel. These computers might figure out what others want by paying attention to how they talk and not just what they say. Making computers self-aware is tricky. First, humans need to understand what it’s like to be aware, and then they have to teach computers to be aware too.
Examples of Artificial Intelligence:
Artificial intelligence comes in different forms, like talking robots, map apps, and fitness trackers you wear. Here are some examples to show how AI can be used:
ChatGPT
ChatGPT is a smart robot that can write things in lots of different ways, like essays or computer code. It was made by OpenAI and works like a human writer. You can even use it on your phone!
Google Maps
Google Maps uses phones and info from people to see how cars are moving on the roads. It can tell you the quickest way to go, taking into account things like traffic and accidents.
Smart Assistants
Think of Hey Google, Siri, Alexa, and Cortana as friendly helpers. They understand what you say and can-do things like remind you of stuff, find info online, and even control your home’s lights. They learn what you like and get better at helping you over time.
Snapchat Filters
Snapchat makes your pictures fun with special effects. It uses clever computer tricks to follow your face and change the picture based on what you’re doing.
Self-Driving Cars
Imagine cars that can drive themselves. They use fancy computer networks to see what’s around them, like other cars and traffic lights. This helps them drive safely without a human driver.
Wearables
Wearable gadgets in healthcare use smart technology to check how healthy you are. They can measure things like your heart rate and blood pressure, and even predict if you might get sick in the future by looking at your medical history.
MuZero
MuZero is a computer program that’s really smart. It can play games like chess and video games without anyone teaching it. It learns by playing lots of times and figuring out how to win. It’s a step closer to super smart AI.
What is Machine Learning and Deep Learning
Although we often hear the words “machine learning” and “deep learning” when we talk about AI, it’s important to know they’re not the same thing. Deep learning is a kind of machine learning, and machine learning is a part of artificial intelligence.
Machine Learning
In machine learning, a computer gets data and uses math tricks to learn and improve at a task. It doesn’t need to be directly taught for that task. Instead, it uses past data to guess what the new results will be. Machine learning has two types: one where the computer knows the right answers beforehand (supervised learning) and one where it doesn’t (unsupervised learning).
Deep Learning
A special kind of machine learning is Deep learning. It uses networks that are like our brain’s nerves. These networks have hidden parts that process data. They let the computer go “deep” in its learning by finding patterns and figuring out what’s important in the data for better outcomes.
Benefits, Future and challenges of AI
Benefits of Artificial Intelligence
Artificial intelligence (AI) is really helpful in many ways, like helping scientists create vaccines faster and catching possible fraud. In 2022, AI companies got a lot of money, $66.8 billion, which is more than double what they got in 2020. People are using AI quickly in different jobs and industries.
Safer Banking Banks are getting safer with AI. More than half of the companies that handle money are already using AI to manage risks and make more money. If they keep using AI, they could save up to $400 billion.
Better Medicine AI is also helping doctors. Even though using AI in healthcare is tricky, it can help a lot. It might make health policies smarter and help doctors find out what’s wrong with patients more accurately.
Cool Entertainment AI is changing how we have fun too. In the future, AI in games, movies, and more could be worth $99.48 billion. That’s a huge jump from $10.87 billion in 2021. AI can do things like catch people who copy others’ work and make super clear graphics.
Future of Artificial Intelligence
Thinking about the money and technology needed for artificial intelligence, actually making AI work is complicated and expensive. But there’s good news! Computers have gotten a lot better, which we can see from Moore’s Law. This rule says that computer parts become twice as powerful every two years, but they also become half as expensive.
Even though experts think Moore’s Law might stop working in the 2020s, it’s helped AI a lot. Without it, AI wouldn’t be as smart. Some new research shows that AI is actually getting even better than Moore’s Law. It’s getting twice as good every six months, not two years.
Because of this, AI has gotten really good in many jobs. And it seems like it will keep getting better in the next few decades.
AI Challenges and Limits
Even though people think AI is really important and getting better quickly, it has its own problems too.
A group called Pew Research Center talked to 10,260 Americans in 2021 about how they feel about AI. The results showed that 45% of them feel both happy and worried about AI, and 37% feel more worried than happy. More than 40% of the people said they think self-driving cars are not good for society. But when it comes to using AI to find fake news on social media, about 40% of the people thought it was a good idea.
AI is great for getting things done faster and better. It can also stop mistakes that people might make. But there are also bad things about it. It can cost a lot to make AI, and machines might take away some jobs from people. But don’t forget, AI can also make new jobs that we don’t even know about yet.
Check out the Timeline of AI: History of Artificial Intelligence
AI History
Robots that could think and fake people started in old stories from Greece. A smart person named Aristotle made a big step by using clever thinking to figure things out. Even though the beginning was a long time ago, the real history of AI is only about 100 years old. Let’s see some important things that happened in AI.
In 1940s
- (1942) Isaac Asimov makes the Three Laws of Robotics, a sci-fi idea saying AI shouldn’t hurt humans.
- (1943) Warren McCullough and Walter Pitts write about a math way to build a brain-like network.
- (1949) Donald Hebb says that our brains get better when we learn from experiences, and this idea still helps AI today.
In 1950s
- (1950) Alan Turing writes about the Turing Test, a way to know if a machine is smart.
- (1950) Students Marvin Minsky and Dean Edmonds create SNARC, a computer that works like a brain.
- (1950) Claude Shannon makes a computer play chess.
- (1952) Arthur Samuel makes a program that learns to play checkers by itself.
- (1954) A machine translates Russian to English.
- (1956) People start using the term “artificial intelligence” at a conference led by John McCarthy.
- (1956) Allen Newell and Herbert Simon make Logic Theorist (LT), a program that thinks.
- (1958) John McCarthy makes a computer language called Lisp and talks about a smart system called Advice Taker.
- (1959) Allen Newell, Herbert Simon, and J.C. Shaw create General Problem Solver (GPS) to solve problems like humans.
- (1959) Herbert Gelernter builds a program that solves math problems.
- (1959) Arthur Samuel uses the term “machine learning” at IBM.
- (1959) John McCarthy and Marvin Minsky start the MIT Artificial Intelligence Project.
In 1960s
- (1963) John McCarthy starts the AI Lab at Stanford University.
- (1966) A report by the U.S. government talks about how machine translations are not going well. This makes them stop funding projects that were supposed to quickly translate Russian.
- (1969) Clever systems like DENDRAL and MYCIN are made at Stanford. They can solve tough problems.
In 1970s
- (1972) People create PROLOG, a language that computers can understand.
- (1973) The British government is not happy with AI research and cuts the funding for projects.
- (1974-1980) AI research gets slow because of money problems. DARPA, a research agency, stops giving out money for AI. This time is called the “First AI Winter.”
In 1980s
- (1980) Digital Equipment Corporation’s makes R1, the first smart computer system for businesses. It starts a trend of smart systems and ends the first AI Winter.
- (1982) Japan’s Ministry of International Trade and Industry begins a big project for powerful computers and AI.
- (1983) The U.S. starts the Strategic Computing Initiative to support advanced computing and AI, because of Japan’s project.
- (1985) Companies spend lots of money on smart systems, and a new market for Lisp computers starts. Companies like Symbolics and Lisp Machines Inc. make special computers for AI.
- (1987-1993) Things change, and the Lisp computer market goes down in 1987. This starts the “Second AI Winter.” Smart systems cost too much to keep up, so people stop liking them.
In 1990s
- (1991) In a war, the U.S. uses DART, a computer tool, to plan things.
- (1992) Japan ends its big computer project in 1992 because it didn’t work well.
- (1993) DARPA stops the Strategic Computing Initiative in 1993 after spending a lot of money, but not getting what they wanted.
- (1997) A computer from IBM named Deep Blue beats the best chess player in the world, Gary Kasparov.
In 2000s
- (2005) STANLEY, a car that drives itself, wins a big race called DARPA Grand Challenge.
- (2005) The U.S. military starts using robots that can think on their own, like “Big Dog” from Boston Dynamics and “PackBot” from iRobot.
- (2008) Google gets better at understanding what people say and puts it in their iPhone app.
In 2010s
- (2011) IBM’s Watson wins on a TV game show called Jeopardy!
- (2011) Apple adds a smart helper named Siri to iPhones.
- (2012) Andrew Ng makes a computer learn from YouTube videos without telling it what things are. This starts a time of smarter computer programs.
- (2014) Google makes a self-driving car that passes a driving test.
- (2014) Amazon’s smart home device, Alexa, comes out.
- (2016) A computer beats a champion player at the ancient game Go, showing how smart computers have become.
- (2016) A robot named Sophia can talk, understand faces, and show feelings.
- (2018) Google makes a tool called BERT that helps computers understand human language better.
- (2018) Waymo lets people in Phoenix get rides in self-driving cars.
In 2020s
- (2020) Baidu makes a computer program that helps scientists work faster on a vaccine for a virus. It’s super-fast at figuring out the virus’s code.
- (2020) OpenAI releases a computer program called GPT-3 that talks and writes like a person.
- (2021) OpenAI makes a computer program named DALL-E that creates pictures from words.
- (2022) A group makes a guide to handle risks with AI, to keep people and society safe.
- (2022) DeepMind shows Gato, a smart computer that can do many things, like play games and move objects.
- (2022) OpenAI creates ChatGPT, a chatbot that many people use.
- (2023) Microsoft makes Bing, a search engine, smarter with AI, like ChatGPT.
- (2023) Google makes Bard, another talking computer.
- (2023) OpenAI makes GPT-4, their best talker yet.
FAQs
Q. Who is the father of AI?
A. John McCarthy is known as the “dad” of Artificial Intelligence. He was a computer expert from the United States. He was the one who made up the words “artificial intelligence.” John McCarthy, along with Alan Turing, Marvin Minsky, Allen Newell, and Herbert A., started the idea of artificial intelligence.
Q. Who is the father of Indian AI?
A. Raj Reddy, whose full name is Dabbala Rajagopal Reddy, was born on June 13, 1937, in a place called Katur (or Katoor) in India. He’s a smart person who knows a lot about computers from India. In 1994, he won a prize along with another smart computer person from America named Edward Feigenbaum.
Q. Who is the first CEO of AI?
A. DICTADOR’s MIKA, the very first humanoid AI CEO in the world, showed up with a lot of strength at the SALZ21-Home of Innovation conference.
Q. What was the first AI device?
A. Back in 1951, a machine called Ferranti Mark 1 learned how to play checkers using a set of instructions. After that, Newell and Simon made a program called General Problem Solver to solve math problems.
