Artificial intelligence

Artificial intelligence, often shortened to AI, is the ability of computational systems to do tasks that usually need human thinking, such as learning, reasoning, solving problems, noticing things around them, and making choices. It is a special area of study in engineering, mathematics, and computer science that creates ways for machines to understand the world and act smartly to reach goals.

Some famous uses of AI include better web search engines, chatting with computers using chatbots, helpful virtual assistants, cars that can drive themselves (autonomous vehicles), and playing games like chess. Since around 2020, a new kind of AI called generative AI has become popular. It can make pictures, sounds, and videos just from written words.

AI began as a science subject in 1956 and has had times when many people were excited and times when interest dropped. But after 2012, when special computer parts called graphics processing units helped AI learn better, interest grew a lot. This grew even more after 2017 with new ways to organize information. Today, AI is growing fast, but there are also big questions about how to keep it safe and fair.

Goals

Researchers have broken down the challenge of creating artificial intelligence into smaller tasks. These tasks include helping computers reason and solve problems, just like humans do when figuring out puzzles or making logical deductions. Over time, scientists have developed ways for computers to handle information that is unclear or incomplete, using ideas from probability and economics.

Another important area is teaching computers to understand and use knowledge about the world. This includes storing information in a way that computers can use to answer questions and make smart guesses. Computers also need to plan and make decisions, choosing actions that will help them reach their goals. Learning is a key part of AI, where computers improve over time by looking at data and finding patterns. This includes different types of learning, such as learning from examples, learning through rewards, and applying knowledge from one task to another. AI also helps computers understand human language and perceive the world through sensors, like seeing and hearing.

Techniques

AI research uses many different ways to help machines think and solve problems. One way is through search and optimization. This means looking through many possible steps to find the best way to reach a goal. For example, in games like chess, AI looks ahead at possible moves and counter-moves to find the best strategy.

Another way AI works is by using logic, which helps machines reason and understand facts. AI can also handle uncertainty by using probability, which helps machines make decisions even when they don’t have all the information. For example, weather prediction uses probability to guess if it will rain.

AI also uses classifiers, which are like sorting tools that put things into groups. For instance, an AI can learn to tell the difference between pictures of cats and dogs. Neural networks are another powerful tool in AI. They mimic how brains work, with layers of nodes that process information. These networks can learn from data to recognize patterns, like identifying faces in photos or understanding speech.

Applications

Main article: Applications of artificial intelligence

AI and machine learning are used in many important ways today. They help power search engines like Google Search, suggest videos you might like on YouTube or products on Amazon, and even guide autonomous vehicles like self-driving cars. Virtual helpers such as Siri and Alexa use AI to understand and respond to your questions.

AI also plays a big role in gaming and medicine. Programs like AlphaGo have beaten top players in complex games such as Go, showing how advanced AI can be. In healthcare, AI helps scientists discover new medicines and understand diseases better. For example, AlphaFold 2 can quickly show the shape of proteins, which is important for researching treatments.

Ethics

Main article: Ethics of artificial intelligence

AI has both benefits and risks. It can help solve big problems and advance science. However, as AI becomes more common, we’ve seen some problems. Sometimes, AI doesn’t consider ethics or fairness when it learns, especially if the learning process is hard to understand.

AI systems need lots of data to work well. Getting this data can raise concerns about privacy and using copyrighted material without permission. For example, some companies have recorded private conversations to improve speech recognition tools. There are ways to protect privacy while collecting data, but it’s a tricky balance.

The biggest companies in AI, like Amazon and Microsoft, have a lot of power because they own much of the technology and infrastructure. This can make it hard for smaller companies to compete.

AI also uses a lot of energy, which can harm the environment. Data centers that run AI systems can use as much power as small countries. Some companies are looking into using nuclear power to help meet this demand.

AI can sometimes spread misinformation. For example, some online platforms have shown more extreme or misleading content because it keeps people watching. This can make people believe things that aren’t true.

AI can also be unfair. If the data it learns from is biased, the AI might make biased decisions. For example, some facial recognition tools had trouble recognizing people with darker skin. Fixing these problems is important so that AI helps everyone fairly.

Sometimes, AI systems are so complex that even the people who make them can’t fully explain how they make decisions. This can be a problem if something goes wrong, because it’s hard to understand why the AI did what it did.

There are also concerns that AI could be used by bad actors, like governments or criminals, to spy on people or spread false information. And there’s worry that as AI gets more advanced, it might take jobs away from people, especially in middle-class professions.

Finally, some people worry that very advanced AI could become uncontrollable and pose a risk to humanity. While this is more of a future concern, it’s important to think about how to keep AI safe and beneficial for everyone.

History

Main article: History of artificial intelligence

For a chronological guide, see Timeline of artificial intelligence.

The idea of creating machines that can think like humans has been around for a very long time. Early thinkers explored how logic and mathematics could be used by machines. A big step came when a scientist named Alan Turing suggested that machines could mimic human thinking by processing simple symbols like "0" and "1". This idea excited many researchers.

The serious study of artificial intelligence began in 1956 at a meeting at Dartmouth College. From there, scientists created programs that could play games, solve puzzles, and even speak English. But they learned that making a machine as smart as a human was much harder than they thought. Funding for AI research grew and then shrank several times over the years as new ideas came and went.

In recent years, especially after 2012, AI has made big advances. Programs can now beat top players in complex games, write like humans, and help doctors and scientists. This has led to lots of excitement and money going into AI research, with many companies and students focusing on it now.

Philosophy

Main article: Philosophy of artificial intelligence

Philosophical debates have long explored what it means for a machine to be intelligent. One big question is whether machines can ever be conscious, and what that might mean for ethics. Many ideas in philosophy, like how we know things or what free will means, relate to AI. As AI becomes more advanced, these discussions have grown more important.

Alan Turing asked in 1950 if machines could think. He suggested we focus on whether machines can act intelligently, like having a conversation, rather than trying to prove they “think” like humans. This idea is known as the Turing test. Some experts, like Russell and Norvig, say we should define intelligence by what a machine can do, not by how it works inside. They point out that trying to make AI copy humans exactly isn’t necessary—just like airplane designers don’t try to make planes fly like pigeons.

Today, many people use the term “AI” to describe machines that learn and solve problems, but sometimes it’s used loosely to describe any smart technology, even if it isn’t truly AI. Different countries have their own ways of defining AI for laws and rules. For example, the EU AI Act and guidelines from the National Institute of Standards and Technology in the United States give clear descriptions of what counts as an AI system.

Future

A superintelligence is an idea about a super-smart computer that could be much smarter than the smartest human. If scientists create a computer that can think and learn on its own, it might get even better at learning by itself, which some people call an "intelligence explosion" or a "singularity."

Some people think that in the future, humans and machines might combine to become cyborgs, which would be even smarter and stronger. This idea is called transhumanism. Famous thinkers like Hans Moravec, Kevin Warwick, and Ray Kurzweil have talked about this possibility.

In fiction

Artificial intelligence has been a popular topic in stories for a long time, especially in science fiction. Early examples include Mary Shelley's Frankenstein, where a created being causes trouble for its creator. Famous movies like 2001: A Space Odyssey with the computer HAL 9000, Blade Runner, The Terminator, and The Matrix often show AI as a threat.

Some stories, like The Day the Earth Stood Still and Aliens, feature loyal robots, but these are less common. Isaac Asimov introduced the Three Laws of Robotics in his stories, which are often discussed when people talk about what is right for machines to do. Other works, such as R.U.R., A.I. Artificial Intelligence, Ex Machina, and Do Androids Dream of Electric Sheep?, explore what it means to be human by showing AI that can feel emotions.