So, What Is AI?

There's no single right answer — but there is a useful way to think about it. Here's our view, plus the best resources for learning more at every level.

How we think about AI

People disagree about what counts as AI because the term describes a spectrum of technologies, not a single thing. A useful way to think about it is to ask three questions about any system:

Does it learn? Can the system improve its performance based on data or experience, without being explicitly reprogrammed? If yes, you're in AI territory.

Does it infer? Can the system draw conclusions from incomplete information, recognise patterns, or make predictions? That's a strong signal of intelligence.

Does it adapt? Can the system handle novel situations it wasn't specifically designed for? The more it can, the more "intelligent" it is.

Using these lenses, we can map everyday technologies onto a spectrum:

Not AI: Fixed rules and formulas

GPS location, spreadsheet formulas, basic calculators. These follow precise instructions — no learning, no inference, no adaptation.

The grey zone: Smart algorithms

Route optimisation, traditional autocorrect, rules-based chatbots. Sophisticated — but following pre-defined logic rather than learning. Newer versions may use AI underneath.

AI: Systems that learn, infer, and adapt

Speech recognition, spam filters, recommendation engines, smart thermostats, autonomous rerouting, generative AI. These systems improve with data and handle novel situations.

The boundaries shift over time. Yesterday's AI breakthrough becomes today's "just software." And the same feature — like autocorrect — can be rebuilt from simple rules into a neural network without the user ever noticing the change.

The important thing isn't getting the definition "right." It's understanding that AI is a spectrum, that it's already woven into your daily life, and that where you draw the line shapes how you think about its implications for your work, your organisation, and society.


Beginner Getting started with AI

New to AI? These are the best starting points — clear, jargon-free, and designed for people who want to understand what AI is and what it means for them.

Course
AI For Everyone
Andrew Ng · Coursera / DeepLearning.AI
A non-technical course designed for anyone who wants to understand what AI can and can't do, how to spot opportunities, and how to navigate AI in their organisation. Widely regarded as the single best starting point.
coursera.org →
Book
Co-Intelligence: Living and Working with AI
Ethan Mollick · 2024
A practical, engaging guide to working alongside AI. Mollick draws on his experience as a Wharton professor to show how AI changes the way we think, work, and create — with honest assessment of both potential and risks.
Find the book →
Newsletter
One Useful Thing
Ethan Mollick · Substack
A consistently excellent newsletter where Mollick tests new AI tools, shares practical insights, and writes with clarity that makes complex topics accessible. Essential reading for staying current.
oneusefulthing.org →
Video
But What Is a Neural Network?
3Blue1Brown · YouTube
Beautifully animated explanation of how neural networks actually work. No prerequisites needed — just curiosity. The best visual explainer of AI's core technology available anywhere.
Watch on YouTube →

Intermediate Going deeper

You understand the basics and want to go further — into strategy, economics, and the real-world implications of AI for organisations and society.

Book
The Turing Trap: The Promise and Peril of Human-Like AI
Erik Brynjolfsson · 2022
Brynjolfsson argues that the obsession with building AI that imitates humans distracts from AI's real potential: augmenting human capabilities in ways we haven't imagined. An essential reframe for leaders and policymakers.
brynjolfsson.com →
Book
Making AI Work for Britain
Alan Brown · London Publishing Partnership, 2026
A practical analysis of where the UK stands with AI — what's working, what isn't, and what choices leaders need to make. Covers governance, skills, public sector adoption, and the gap between AI ambition and implementation reality.
More info →
Report
AI Index Report
Stanford HAI · Annual
The most comprehensive annual survey of AI progress — covering research, industry, policy, education, and public opinion with extensive data and visualisations. The reference document for anyone who needs facts, not hype.
aiindex.stanford.edu →
Course
AI for Business Specialization
Andrew Ng · Coursera / DeepLearning.AI
The next step after AI For Everyone. Covers how to build an AI strategy, evaluate AI projects, work with AI teams, and navigate the organisational challenges of implementation.
deeplearning.ai →

Advanced For practitioners and researchers

You're building with AI or shaping AI strategy, and you need deep, current, technically grounded material.

Course
Deep Learning Specialization
Andrew Ng · Coursera / DeepLearning.AI
Five courses that take you from the foundations of neural networks through to sequence models and attention mechanisms. The gold standard for understanding how modern AI systems actually work under the hood.
coursera.org →
Research
Attention Is All You Need
Vaswani et al. · Google, 2017
The paper that introduced the Transformer architecture — the foundation of GPT, Claude, Gemini, and virtually every modern large language model. If you want to understand the technical basis of the current AI wave, this is where it starts.
arxiv.org →
Research
NBER Working Papers on AI Economics
Erik Brynjolfsson, Daron Acemoglu, and others
Rigorous research on AI's economic impact — productivity effects, labour market disruption, inequality, and the conditions under which AI creates or destroys value. Essential reading for anyone making policy or investment decisions about AI.
nber.org →
Newsletter
Import AI
Jack Clark
Weekly newsletter covering the latest AI research, policy developments, and industry moves. Written by one of OpenAI's co-founders, it strikes the rare balance between technical depth and strategic insight.
importai.substack.com →