What is Artificial Intelligence?
Artificial intelligence, or AI, is the branch of computer science that seeks to create systems capable of performing tasks which, until recently, required human intelligence: understanding language, recognising images, making decisions or learning from experience.
Until very recently, this was the stuff of science fiction. But it is now part of our daily lives: when your phone unlocks using facial recognition, when Netflix recommends a series, when your email filters out spam, or when you ask a virtual assistant a question. AI works quietly, but its impact is becoming increasingly visible.
The truth is that AI does not mimic the human brain: it learns from data on a scale and at a speed that no human could match.
A bit of history
The term was coined quite some years ago, in 1956, by the mathematician John McCarthy during a conference at Dartmouth. For decades, AI progressed slowly, held back by a lack of data and computing power. The big leap came in the second decade of the 21st century, when the explosion of big data, graphics processing units (GPUs) and deep learning algorithms made it possible to train models of a complexity that was previously unthinkable.
Today, language models such as those powering ChatGPT or Claude have been trained on hundreds of billions of words, enabling them to hold conversations, write code, summarise documents or generate creative content with surprising fluency.
Types of AI
- Narrow AI: Designed for a specific task. It plays chess, recognises faces or translates text. This is the AI that exists today.
- General AI (AGI): Hypothetical. It could reason and learn in any domain just as a human would. It does not yet exist.
- Superintelligence: Theoretical. An AI that surpasses humanity in all cognitive abilities. A subject of philosophical debate.
But how does AI learn?
The dominant technique today is called machine learning. Instead of programming fixed rules, it is shown thousands or millions of examples, and the system internally adjusts its parameters until it can predict or classify correctly.
Within machine learning, deep learning uses artificial neural networks — layers of mathematical nodes loosely inspired by biological neurons — which enable the detection of complex patterns in images, audio or text.
Opportunities and risks
AI promises to accelerate medical research, combat climate change, democratise access to knowledge and boost productivity across millions of sectors. However, it also raises serious questions: algorithmic bias, the impact on employment, data privacy, automatically generated misinformation, and the concentration of technological power in a few hands.
The challenge facing our generation is not merely to develop more powerful AI, but AI that is fairer, more transparent and aligned with human values. Organisations such as Anthropic, DeepMind and the Centre for AI Safety are working precisely towards this goal.
Ultimately, artificial intelligence is neither a magic solution nor an inevitable threat. It is an extraordinarily powerful tool that amplifies what we already are. But the question is not whether AI will transform the world. It already is. The question is how we want it to do so.