What Is AI? An Introduction to the Basic Types and Use Cases

The term “artificial intelligence” was first popularized at the 1956 Dartmouth Conferences, and until the past few years it was largely considered science fiction. From R2D2 and C3PO to The Terminator, people have always wondered what it would be like if machines could think, learn, reason and behave like humans. AI made for great entertainment, but it was never realistic.

Now that more compute power is available, AI is no longer far-fetched. Graphics processing units (GPUs), which perform calculations much faster than traditional central processing units (CPUs), have helped enable the rapid growth of AI applications. Demand for AI hardware is high — TrendForce predicts that about 1.2 million AI servers will ship in 2023, representing almost 9 percent of total server shipments.

Today, it’s within the reach of almost any business to invest in AI technology. Before making those investments, however, organizations need to understand what AI is, what it is not, and how it can be used.

Interested In Learning More?

What Is AI?

At the most basic level, AI involves the development of machines that can simulate human intelligence. AI systems are capable of performing tasks that require cognitive skills. These systems fall into two broad categories:

Weak AI, or narrow AI, applies intelligence to a specific problem or task. While it may perform that task very well, it is far more limited than basic human intelligence.
Strong AI, or artificial general intelligence, is capable of solving any kind of problem, like a human. This type of artificial intelligence does not yet exist outside the world of science fiction.

Weak AI systems are getting stronger. Technological advances in the past ten years have made AI part of everyday life and a common feature in many businesses. The five categories of AI tools are as follows:

Chatbots and virtual assistants, such as Siri, Alexa and customer service chatbots
Image and facial recognition systems, including those used for user authentication on many smartphones
Predictive maintenance systems that use data from sensors to alert users of potential equipment failure
Recommendation engines that determine content a user might like on social media and streaming media platforms
Self-driving vehicles, including cars, industrial robots and drones

How Does AI Work?

AI systems are programmed with algorithms that enable them to look for patterns in data. The goal is to simulate learning and reasoning. AI systems must learn how to analyze data and make decisions based on the results. They must also determine the right method for achieving the desired objective, and continually fine-tune their processes to improve results. The most advanced AI systems can generate new text, images, software or music based on statistical probabilities.

Different types of AI systems learn in different ways:

Machine learning uses statistical techniques to solve problems without being programmed to do so. The machine learning algorithm is trained using labeled data so that it can begin to identify underlying patterns. The trained model can then make predictions on new, unlabeled data.
Deep learning is an advanced form of machine learning that uses multilayered neural networks to mimic the activity of neurons in the human brain. Data passes through multiple layers of processing to identify subtle correlations and produce a reliable output. Deep learning requires very large amounts of training data, which can be labeled or unlabeled.
Natural language processing enables computers to understand spoken or written language. NLP combines machine learning, deep learning and statistical models with computational linguistics. Speech-to-text software, chatbots and translation programs are examples of NLP.
Computer vision gives computers the ability to understand images, videos and other visual inputs. The system must learn how to distinguish different objects, determine how far away they are and more. Once it’s successfully trained, a computer vision system can analyze visual inputs much faster than humans.

The Path Forward

To take advantage of AI, an organization must first identify the problem to be solved. This will drive the type of AI system to be employed. In our next series of posts, we’ll take a deeper dive into deep learning, NLP and computer vision. We’ll explore use cases for these technologies and some of the risks that organizations need to consider.