Generative AI outputs new content, not verified facts

Generative AI stands apart from older AI tools because it produces original content rather than just classifying, sorting, or predicting based on existing data. For example, a traditional AI might label a photo as 'cat' or 'dog', whereas generative AI can create a brand-new image of a cat wearing a hat. The key distinction is creation versus recognition — but creation does not guarantee correctness, and the output can be plausible yet completely wrong.

In an exam, watch for questions that pair 'generates new content' with 'always accurate' — that is a trap. A useful trick is to remember that generative models are trained on patterns, not facts, so they can invent references or produce biased material. To eliminate wrong options, look for verbs like 'creates', 'generates', or 'produces' versus 'analyses', 'classifies', or 'predicts'. If a question mentions 'new content' but also claims the output is 'verified' or 'reliable', that is a red flag.

A compact way to test yourself: ask 'Does this AI make something new, or just sort what already exists?' If it makes something new, it is generative — but never assume the output is true. Pair the phrase 'generates, not verifies' with every generative AI example you see.