4 Types of AI: Getting to Know Artificial Intelligence

Artificial intelligence (AI) technology has created opportunities to progress on real-world problems concerning health, education, and the environment. In some cases, artificial intelligence can do things more efficiently or methodically than human intelligence. 

“Smart” buildings, vehicles, and other technologies can decrease carbon emissions and support people with disabilities. Machine learning, a subset of AI, has enabled engineers to build robots and self-driving cars, recognize speech and images, and forecast market trends. 

Read on to learn more about the four main types of AI—reactive machines, limited memory machines, theory of mind, and self-awareness—and their functions in everyday life.

4 main types of artificial intelligence

Learning in AI can fall under the types “narrow intelligence,” “artificial general intelligence,” and “super AI.” These categories demonstrate AI’s capabilities as it evolves—performing narrowly defined sets of tasks, simulating thought processes in the human mind, and performing beyond human capability.

Watch this video from IBM's Introduction to Artificial Intelligence (AI) course to learn more about these categories:

There are four main types of AI as defined by Arend Hintze, researcher and professor of integrative biology at Michigan State University [1]. They are as follows:

1. Reactive machines

Reactive machines are AI systems that have no memory and are task-specific, meaning that an input always delivers the same output. Machine learning models tend to be reactive machines because they take customer data, such as purchase or search history, and use it to deliver recommendations to the same customers.  

This type of AI is reactive. It performs “super” AI because the average human would not be able to process huge amounts of data, such as a customer’s entire Netflix history and feedback customized recommendations. Reactive AI, for the most part, is reliable and works well in inventions like self-driving cars. It doesn’t have the ability to predict future outcomes unless it has been fed the appropriate information.

Compare this to our human lives, where most of our actions are not reactive because we don’t have all the information we need to react upon, but we have the capability to remember and learn. Based on those successes or failures, we may act differently in the future if faced with a similar situation.

Examples of reactive machines

Beat at chess by IBM’s supercomputer: One of the best examples of reactive AI is when Deep Blue, IBM’s chess-playing AI system, beat Garry Kasparov in the late 1990s. Deep Blue could identify its own and its opponent’s pieces on the chessboard to make predictions, but it does not have the memory capacity to use past mistakes to inform future decisions. It only makes predictions based on what moves could be next for both players and selects the best move. 

Netflix recommendations: Netflix’s recommendation engine is powered by machine learning models that process the data collected from a customer’s viewing history to determine specific movies and TV shows that they will enjoy. Humans are creatures of habit—if someone tends to watch a lot of Korean dramas, Netflix will show a preview of new releases on the home page.

2. Limited memory machines

The next type of AI in its evolution is limited memory. This algorithm imitates the way our brains’ neurons work together, meaning that it gets smarter as it receives more data to train on. Deep learning algorithms improve natural language processing (NLP), image recognition, and other types of reinforcement learning.

Read more: What is Deep Learning? Definition, Examples, and Careers

Limited memory AI, unlike reactive machines, can look into the past and monitor specific objects or situations over time. Then, these observations are programmed into the AI so that its actions can be performed based on both past and present moment data. But in limited memory, this data isn’t saved into the AI’s memory as experience to learn from, the way humans might derive meaning from their successes and failures. The AI improves over time as it’s trained on more data.

Example of limited memory artificial intelligence

Self-driving cars: A good example of limited memory AI is the way self-driving cars observe other cars on the road for their speed, direction, and proximity. This information is programmed as the car’s representation of the world, such as knowing traffic lights, signs, curves, and bumps in the road. The data helps the car decide when to change lanes so that it does not get hit or cut off by another driver. 

3. Theory of mind

The first two types of AI, reactive machines and limited memory, are types that currently exist. Theory of mind and self-aware AI are theoretical types that could be built in the future. As such, there aren’t any real-world examples yet.

If it is developed, theory of mind AI could have the potential to understand the world and how other entities have thoughts and emotions. In turn, this affects how they behave in relation to those around them.

Human cognitive abilities are capable of processing how our own thoughts and emotions affect others and how others’ affect us—this is the basis of our society’s human relationships. In the future, theory of mind AI machines could be able to understand intentions and predict behavior, as if to simulate human relationships.

What type of AI is ChatGPT?

ChatGPT is an example of generative AI, a kind of artificial intelligence powered by large language models (LLMs) that are created by training algorithms on massive amounts of data. Generative AI is capable of creating original outputs in response to user inputs or "prompts." ChatGPT's ability to produce responses resembling human language has made it a common type of AI used for chatbots and virtual assistants.

That said, while tools like ChatGPT, Google Gemini, or Microsoft Copilot may produce responses that give the impression of a self-aware AI, they aren't really. Instead, their responses are actually just the result of the algorithm identifying the statistically most likely output based on its training data and the user's prompt.

Read more: What Is ChatGPT? (and How to Use It)

4. Self-awareness

The grand finale for the evolution of AI would be to design systems that have a sense of self, a conscious understanding of their existence. This type of AI does not exist yet.

This goes a step beyond theory of mind AI and understanding emotions to being aware of themselves, their state of being, and being able to sense or predict others’ feelings. For example, “I’m hungry” becomes “I know I am hungry” or “I want to eat lasagna because it’s my favorite food.” 

Artificial intelligence and machine learning algorithms are a long way from self-awareness because there is still so much to uncover about the human brain’s intelligence and how memory, learning, and decision-making work.

Learn more about artificial intelligence on Coursera

Learning about AI can be fun and fascinating, even if you don’t want to become an AI engineer. Whether you just want to boost your productivity in the workplace or build toward a career in AI, Coursera has something for you:

For an overview of AI, take DeepLearning.AI's AI for Everyone course. Designed for non-technical people to understand what AI is, the course includes common terminology like neural networks, machine learning, deep learning, and data science. You’ll learn how to work with an AI team build an AI strategy in your company, and much more. 

To leverage the potential of generative AI, consider enrolling in the IBM Generative AI Fundamentals Specialization. Learn fundamental concepts, models, tools, and AI applications to boost productivity at work and in your life.

To enhance your existing AI skills, explore the IBM Deep Learning with PyTorch, Keras and Tensorflow Professional Certificate. Learn to build, train, and deploy deep learning models in this intermediate-level program.