Episode 40: What is AI?
- Embedded IT
- May 26, 2025
- 3 min read
Updated: Jan 16
Artificial intelligence has been getting a lot of attention, but it is far from new. This episode breaks down what AI actually is, how it has evolved, and the layers that sit beneath it. The aim is to bring clarity to a topic that often gets overhyped, especially for anyone involved in technology procurement.
If you’re coming to AI for the first time, start with our intro to the AI series to see what we cover and why it matters.
The early history of artificial intelligence
AI has roots going back to the 1950s. A key milestone was the creation of the Turing Test, designed to assess whether a computer could appear as intelligent as a human. The idea was simple. A person would ask questions to two rooms, one containing a human and one containing a computer. If the person could not tell which answers were computer generated, the machine had effectively passed the test.
For decades, computers were nowhere near capable of achieving this consistently. They were not powerful enough and were largely focused on pure mathematical tasks.
The rise of neural networks
Things began to shift in the 1980s with the development of neural networks. These systems simulate the way the human brain works by connecting data points to predict outcomes. While the detail is complex, the principle is straightforward. Neural networks made it possible for computers to learn in a more sophisticated way.
A major milestone in the 1990s
In the early 1990s, IBM created an AI system designed to play chess at a world-leading level. It analysed huge numbers of potential moves to identify the strongest option. The system famously defeated chess champion Garry Kasparov, which proved just how advanced computer intelligence had become. This breakthrough pushed AI firmly into public view.
The acceleration of modern AI
Fast forward to today and AI has developed at an incredible pace. More powerful computers, huge amounts of available data, and digital systems everywhere have created the perfect environment for rapid advancement. Organisations such as OpenAI have helped bring AI into the mainstream, with tools that demonstrate just how capable generative systems can be.
The three layers of AI
AI is often described in three layers that build on each other:
Machine learning
The foundation layer. A computer is fed large amounts of data so it can learn patterns and predict likely outcomes. It is basic but effective.
Deep learning
A more advanced layer built on neural networks. Deep learning considers multiple variables at once, analysing how they interact to reach a conclusion.
Generative AI
The current headline technology. Generative AI links large volumes of data, learns from it, and then produces entirely new content. Large language models, such as those behind modern chat tools, sit within this category.
Why this matters for procurement and society
AI is not new, but it is moving fast. With more data and more powerful systems than ever before, AI is set to transform how organisations work and how society operates. Future episodes will explore what this means in practical terms, including the impact on procurement decisions.
To get guidance on how AI will influence future technology procurement, get in touch.
