AI Fundamentals

Artificial Intelligence

Artificial intelligence, or AI, is the broad field concerned with creating computer systems capable of performing tasks that normally require aspects of human intelligence.

These tasks can include understanding language, recognising images, identifying patterns, making predictions, solving problems, planning actions and generating new content.

AI is an umbrella term rather than one specific technology. Machine learning, deep learning, generative AI and large language models are all areas or technologies that sit within the wider field of artificial intelligence.

Machine Learning

Machine learning is a branch of artificial intelligence in which computer systems learn patterns from data rather than relying entirely on explicitly programmed rules.

A machine-learning model is typically trained using examples and then uses what it has learned to make predictions, classifications or decisions when presented with new information.

Machine learning underpins a large proportion of modern AI, from recommendation systems and fraud detection to image recognition and language models.

Deep Learning

Deep learning is a form of machine learning built around neural networks containing multiple layers of computation.

These networks can learn complex patterns from large amounts of data and have played a central role in advances across computer vision, speech recognition and generative AI.

The rise of deep learning helped make it possible to train increasingly capable models without manually defining every feature the system should look for.

Neural Networks

A neural network is a computational system made up of interconnected processing units that transform information through a series of weighted connections.

Neural networks are loosely inspired by the structure of biological neural systems, although modern artificial neural networks should not be understood as direct simulations of the human brain.

They form the technological foundation of deep learning and are used in many contemporary AI systems.

Generative AI

Generative AI refers to artificial intelligence systems designed to create new content, including text, images, audio, video, computer code and other forms of data.

Rather than simply classifying or analysing information, generative models learn patterns within training data and use those patterns to produce new outputs.

Large language models are one major form of generative AI, while diffusion models have become particularly important in image and video generation.

Large Language Models

A large language model, or LLM, is an AI model trained on large amounts of text and other data to understand and generate language.

LLMs work by modelling relationships between tokens and predicting appropriate continuations based on the information available within their context.

They power many contemporary AI assistants, coding tools, search systems and agents, but the term describes the underlying model rather than the complete application built around it.

Foundation Models

A foundation model is a large AI model trained on broad datasets that can subsequently be adapted or used across many different tasks.

Instead of being developed for one narrow purpose, foundation models can provide the underlying intelligence for numerous applications.

Many leading large language models are foundation models, but foundation models can also work with images, audio, video and other forms of information.

Transformers

A transformer is a neural-network architecture that has become fundamental to modern language models and many other AI systems.

Introduced in 2017, the transformer architecture uses mechanisms known as attention to determine which pieces of information are most relevant to one another.

Transformers made it practical to train models on enormous quantities of sequential data and became the foundation for most of the large language models used today.

Natural Language Processing

Natural language processing, or NLP, is the field of artificial intelligence concerned with enabling computers to work with human language.

NLP includes tasks such as translation, summarisation, sentiment analysis, information extraction, speech processing and text generation.

Large language models have dramatically changed the capabilities of NLP systems, but NLP itself predates the current generation of generative AI by decades.

Computer Vision

Computer vision is the field of artificial intelligence focused on enabling machines to analyse and understand images and video.

Computer-vision systems can be used for tasks including object detection, facial recognition, medical imaging, industrial inspection, autonomous driving and satellite analysis.

Modern multimodal AI increasingly combines computer vision with language and other types of information.

Multimodal AI

Multimodal AI refers to artificial intelligence systems capable of processing or generating more than one type of information.

A multimodal model might work with combinations of text, images, audio, video or other data rather than being restricted to a single input type.

This allows AI systems to interpret information more like a connected environment rather than treating language, images and sound as completely separate problems.

Artificial General Intelligence

Artificial general intelligence, usually abbreviated to AGI, describes a hypothetical form of AI capable of performing or learning across a very broad range of intellectual tasks rather than being specialised for particular applications.

There is no universally accepted technical definition of AGI and no agreed threshold at which an AI system should be considered generally intelligent.

The term nevertheless plays an important role in AI research and strategy, particularly among organisations attempting to build systems with increasingly general reasoning, learning and problem-solving abilities.

Artificial Superintelligence

Artificial superintelligence, or ASI, refers to a hypothetical form of artificial intelligence whose capabilities substantially exceed those of humans across most or all intellectual domains.

ASI remains a theoretical concept rather than an established technological category.

It is nevertheless central to discussions around the long-term development of AI, particularly in relation to alignment, safety, control and the possibility of increasingly capable systems improving themselves.