What Is AI?


Artificial Intelligence

Artificial intelligence (AI) is an interdisciplinary field combining mathematics, statistics, cognitive science, and computing to solve complex problems using large datasets and high-performance computers.

Within AI, machine learning and deep learning are key subfields. Machine learning involves algorithms that enable systems to learn from data, while deep learning, the most advanced form, closely mimics human intelligence using neural networks.

Examples of AI applications include:

  • Self-driving cars
  • Speech and facial recognition
  • Digital personal assistants
  • Chatbots and virtual customer service
  • Recommendation engines

illustration of what is ai subfields

Is it AI or computer science?  

Artificial intelligence (AI) incorporates various fields such as computer science, mathematics, philosophy, biology, psychology, and neuroscience. Its applications in research are broad, ranging from infrastructure development and construction to drug discovery, medical diagnosis, and environmental protection. At LSU, researchers across multiple disciplines are leveraging AI to advance their work in diverse fields, applying AI as a powerful tool for problem-solving and innovation.

How does it work?

deep learning illustration

Artificial intelligence (AI) can learn through trial and error, just like humans and animals. When a machine makes a mistake, it learns from it and improves the next time it faces the same problem. AI uses machine learning algorithms to break data into smaller parts to understand new information. There are two types of learning: supervised and unsupervised.

In supervised learning, the AI is given labeled data. For example, if you show it pictures of cats and dogs, at first, it won’t know which is which, so humans help by labeling the images ("this is a cat," "this is a dog"). Over time, the AI learns to recognize the differences on its own, such as cats having whiskers.

Unsupervised learning doesn’t need labeled data, so the machine figures things out on its own, which is becoming more popular since it’s harder to get labeled data.

Deep learning is a type of AI that uses layers of artificial neurons, similar to how the human brain works. Each layer of these networks learns more complex features from the data, like recognizing lines, shapes, and then more detailed things like animals. Deep learning is key in fields like computer vision, which allows AI to "see" the world—important for things like facial recognition and self-driving cars. One big challenge for AI is perception, like recognizing a stop sign that’s partially hidden by shadows, which remains tricky for AI systems to handle correctly.

Is AI safe? (The Terminator question)

Current deep learning systems often don’t clearly explain how they make decisions, which can be a problem when the data they use isn’t very accurate. This can happen when the data is noisy or incomplete, and it makes the system’s decisions less reliable. In some cases, attackers might even take advantage of this weakness.

Ideally, AI systems should also explain how confident they are in their decisions, so people can better judge if the outcome is trustworthy. LSU researcher Supratik Mukhopadhyay is working on solving this problem by developing "confidence-aware AI," which helps make AI decisions safer and more transparent.

Why AI matters

AI, in simple terms, helps us make better decisions by quickly analyzing huge amounts of data, much faster than a human could. It’s really good at spotting patterns, even ones we might not know to look for. AI is already being used for things like climate change research, improving cancer treatments, and designing energy-efficient buildings. As we keep using AI, it will continue to help us make smarter choices and better predictions in many areas.