What is Generative Ai?
What Is Generative AI?
Generative artificial intelligence is a type of AI that can create new content and ideas, including conversations, stories, images, videos and music. Explore its key language, common uses, and the machine-learning engine underneath it.
Write a two-line birthday message.
Wishing you a joyful birthday filled with bright moments and happy memories!
Key Terms, Indexed
Six terms form the basic language of generative AI. Select a term to reveal its definition, relationship and example. Understanding the difference between the user's prompt and the model's completion is especially important.
How the Fields Fit Together
These concepts can be viewed as nested fields. Artificial intelligence is the broadest field. Machine learning sits within AI, deep learning sits within machine learning, and generative AI is a specialised application built using deep-learning models.
Uses of Generative AI
Generative AI can support personal and organisational work, answer queries, create and summarise content, translate between text and images, follow instructions, and help write computer programs. Select each use to see what it does and how it appears in practice.
| Ref. | Use | Typical output | Human responsibility |
|---|---|---|---|
| 1.2.a | Personal or organisational use | Ideas, drafts or workflow support | Apply suitable organisational controls. |
| 1.2.b | Responding to queries | Direct natural-language answers | Check important facts. |
| 1.2.c | Content creation | Text, images and other new content | Review and edit before use. |
| 1.2.d | Summarising documents | Condensed key points | Confirm that important meaning remains. |
| 1.2.e | Text ↔ image | Generated images or descriptions | Check suitability and accuracy. |
| 1.2.f | Following instructions | Completed multi-step tasks | Supervise actions and permissions. |
| 1.2.g | Writing programs | Draft, explained or revised code | Review and test the code. |
The Machine-Learning Engine
Generative AI is built on machine learning and, more specifically, deep learning. A model is produced through a repeatable process: define the problem, prepare suitable data, select and improve a model, and review whether the outcome meets the original goal.
Machine Learning
The study of computer algorithms that allow computer programs to automatically improve through experience. Instead of relying only on fixed, hand-written rules, the program learns patterns from data.
Deep Learning
A multi-layered neural network in which each layer extracts a more abstract representation of the data than the layer before it.
Select a stage to follow the six-stage machine-learning process.
Remember the Distinctions
Umbrella Field
Machine intelligence is the broad field containing machine learning and deep learning.
Creates Content
Deep-learning models generate text, images and other content from learned data.
Language Model
An LLM recognises, summarises, translates, predicts and generates content using large datasets.
Human Language
NLP enables computer programs to understand spoken and written language.
Input and Output
The prompt is given to the model; the completion is generated in response.
Repeat and Review
A model is trained, tested and improved within a wider problem-solving process.