What is AI?
Definition of AI
In 1955, emeritus Stanford Professor John McCarthy coined the term “artificial intelligence” (AI) and defined it as “the science and engineering of making intelligent machines.”
Today, the National Artificial Intelligence Act of 2020 defines AI as “a machine-based system that can, for a given set of human-defined objectives, make predictions, recommendations or decisions influencing real or virtual environments.”
The next level of AI, generative AI, goes beyond traditional AI capabilities of data analysis, specific tasks, and predictions. Generative AI creates new content (text, images that is similar to its training data.
How does AI Work?
AI algorithms extract data patterns and make predictions or decisions using large datasets that allow the software to learn from the patterns in the data without explicit programming. This enables it to reason, solve problems, and analyze human language (natural language processing). AI along with sense captors can also interpret and understand visual data, image or character recognition, facial recognition, object detection, temperature, and other perceptive inputs.
Types of AI
AI includes many subsets and they can be based on capabilities or functionalities.
| Type | Existence | Capabilities | Examples | Notes |
|---|---|---|---|---|
| Artificial Narrow AI | Practical | Can perform a single or narrow task often better or faster than a human mind. | Siri, Alexa, ChatGPT | The only type of AI that exists today. |
| General AI | Theoretical | Can use previous learnings and skills to accomplish new tasks in a different context without human beings to train the underlying models. | N/A | Also known as "strong AI". |
| Super AI | Theoretical | Can think, reason, learn, make judgements, and possess cognitive abilities that surpass human beings. | N/A | Also known as "artificial superintelligence". |
Adapted from IBM types of artificial intelligence article
| Type Based on Functionality | Type Based on Capability | Functionality | Examples |
|---|---|---|---|
| Reactive Machine AI | Narrow | Task oriented with no memory and no access to previous outcomes or decisions, only work with presently available data. | Netflix Recommendation Engine, IBM Deep Blue |
| Limited Memory AI | Narrow | Can recall past events and outcomes and monitor objects or situations over time. Can use past and present data to make decisions. While it doesn't retain past data to use over a long-term period, it can improve in performance over time. | Generative AI such as ChatGPT, Gemini, DeepAI Virtual assistants and chatbots: Siri, Alexa, Google Assistant, Self-driving cars |
| Theory of Mind AI | General AI | Understand thoughts and emotions of other entities which affects how the AI interacts with those around them. It could personalize interactions with individuals based on their unique needs. It could contextualize. Emotion AI is a theory of mind AI currently in development to analyze voices, images and other data to respond on an emotional level. | Theoretical |
| Self-Aware AI | Super AI | Could understand its own internal conditions and traits, introspection, emotions, thoughts, needs, beliefs. | Theoretical |
Adapted from IBM types of artificial intelligence article