Generative AI
AI systems that create new content — text, images, audio, video, code — by learning patterns from training data.
Definition
Generative AI refers to models that produce original content rather than just classifying or analysing inputs. The output can be text, images, audio, video, 3D objects, or code. Unlike discriminative models (which learn decision boundaries), generative models learn the underlying data distribution and can sample from it.
The dominant generative architectures are Transformers (for text/code) and Diffusion Models (for images/video). Generative Adversarial Networks (GANs) and Variational Autoencoders (VAEs) are earlier approaches still used in some applications.
The generative AI market exploded in 2022-2023 with ChatGPT, DALL-E, and Stable Diffusion reaching mainstream adoption. Capabilities now include photorealistic images, coherent long-form text, working code, realistic video clips, and voice cloning.
Examples
- ChatGPT
- DALL-E 3
- Midjourney
- Stable Diffusion
- Sora
- GitHub Copilot
Want a deeper dive?
Read our full explainer with use cases, how-it-works, and FAQs.
Generative AI concept guideCompanies using this
Related Terms
Large Language Model (LLM)
A transformer-based AI system trained on billions of tokens of text, capable of generating, reasoning about, and transforming language.
Diffusion Model
A generative model that learns to create data by reversing a gradual noise-addition process.
GAN (Generative Adversarial Network)
A generative model architecture where two networks — a generator and discriminator — compete to produce realistic synthetic data.
Multimodal AI
AI systems that process and generate multiple data types — text, images, audio, and video — simultaneously.