What Is GenAI?

An introduction to genAI, its principles, and examples of its applications.

GenAI is a remarkable branch of artificial intelligence that enables users to create novel outputs that closely resemble human-generated content. These outputs can take various forms, including text, images, videos, sounds and 3D models. Thanks to recent advancements in the field, genAI has witnessed unprecedented levels of growth and adoption and is revolutionising numerous industries and domains.

The crux of genAI is in its name–generative. While traditional AI algorithms have focused on identifying patterns within data for predictive purposes, say to predict whether the next image in a sequence of images is a cat or a dog, genAI leverages the learned patterns to create entirely new outputs. Given the above example, a genAI model could create an entirely new representation of cat or dog, or cat-dog!

There are now hundreds of genAI platforms with a diverse range of use cases, but undoubtably the most relevant and influential of these is ChatGPT. ChatGPT is a conversational agent that can engage in natural language conversations with humans. It offers a range of applications in higher education and in design, but is often misunderstood as a direct question-and-answer tool. This perception tends to oversimplify ChatGPT and overlooks its potential application in a wider range of more intricate and nuanced tasks.

How does Generative AI work?

GenAI is grounded in machine learning techniques that draw inspiration from the neural systems of the human brain, known as neural networks. These genAI networks are ‘trained’ on extremely large amounts of data, from which they learn to capture and identify features, patterns and relationships within the data. This enables the model to generate new data instances that are similar yet distinct from the original training data.

For instance, a genAI model trained on images of faces learns to understand facial features like nose shape, eye placement, and smile curvature, allowing it to generate new faces that possess the same structure and features but do not match any specific face. Similarly, models like ChatGPT, trained on vast amounts of language data, learn to understand grammar, sentence structure, common phrases, context, tone and style, enabling them to generate coherent and contextually appropriate text.

GenAI can utilise a variety of models, each employing unique methods for training the AI and generating results. There are many types of models, the most popular being Generative Adversarial Networks (GANs), Variational AutoEncoders (VAEs) and Denoising Diffusion Probabilistic Models (DDPMs). Each model possesses unique strengths and limitations, making them suitable for different contexts. Some models excel at producing high-quality results, while others offer better control over the generation process. Consequently, the choice of model plays a crucial role in determining the capabilities and limitations of genAI applications across disciplines.

GenAI, the models used, its applications and its outputs are developing very quickly. Current debates relating to its legal status, control and commercialisation, as well as its potential, will impact the ways it develops into the future and the impact on built environment disciplines and on learning and teaching.