AI has been around much longer than ChatGPT. In 1956, researchers met at Dartmouth College to explore whether machines could do things we associate with human intelligence. Early AI often followed rules written by people. Later, machine learning helped computers find patterns in examples. Today’s AI can recognise speech and images, answer questions and create new content.
There are three broad forms when we talk about how capable AI is:
• Narrow AI does particular tasks, such as recommending a video, recognising a face or helping you write. The AI tools we use today fall in this broad category, even when they can do several tasks.
• Artificial general intelligence (AGI) would learn and handle a wide range of tasks at a human-like level. Researchers are still debating how to define and achieve it.
• Artificial superintelligence (ASI) would be more capable than humans across many areas. This remains hypothetical.
You may also hear names based on what AI does: generative AI creates text, images or audio; predictive AI estimates what may happen; and computer vision helps machines understand pictures. These are uses of AI, not separate stages of intelligence.
Want to read more? See Dartmouth’s history of AI: https://ai.dartmouth.edu/our-story and IBM’s explanation of AI types: https://www.ibm.com/think/topics/artificial-intelligence-types