Since the release of ChatGPT in 2022, Artificial Intelligence (AI) is on everyone's lips. However, you may be surprised to know that AI isn’t a new idea. It’s been developing quietly in the background for more than seventy years. 

There is much speculation about the transformative effects that AI will have on our daily lives, but to truly understand where we are heading, it makes sense to consider where we have come from. In this post, we discuss the history of AI - from the early days in the mid-20th Century, to LLMs today, as well as what we can expect in the future.

The Early Days (1950s–1970s)

The story begins in the 1950s. Computers were the size of rooms, and a small group of mathematicians started asking a bold question: could machines think?

Alan Turing — a British computer pioneer — first proposed a test for “machine intelligence.” In 1956, researchers gathered at Dartmouth College in the US and officially coined the term Artificial Intelligence.

Back then, the goal was to teach machines to reason like humans by writing lots of “if this then that” rules. It worked for very narrow problems, but computers were far too weak to handle anything complex. Governments and investors lost patience, and funding dried up. The first “AI winter” arrived.

Machine Learning Arrives (1980s–2000s)

By the 1980s, the idea returned — this time using data instead of fixed rules. Rather than telling a computer how to solve a problem, programmers started feeding it examples so it could learn patterns for itself.

This approach, called machine learning, produced results in areas like credit scoring, fraud detection and early speech recognition. As computers got faster and cheaper in the 1990s and 2000s, progress accelerated. Businesses began using AI quietly in the background — for recommendations, logistics, and customer analytics.

The Deep Learning Boom (2010s–2020s)

Around 2012 a new method, deep learning, made AI jump forward again. It used “neural networks” — loose digital copies of how the brain processes information.

These systems could recognise photos, translate languages, and even drive cars. The more data and computing power they were given, the better they became.

Then in 2017, researchers created a new type of neural network called the Transformer. It could read and generate human-like text by predicting the next word in a sentence. That breakthrough led to large language models (LLMs) — the engines behind tools like ChatGPT, Copilot and Gemini.

By 2025, these models can understand text, images, and even voice in real time. They’ve moved from research labs to everyday business tools.

AI vs the Human Brain

It helps to keep perspective. The human brain has around 86 billion neurons — tiny cells that connect through trillions of pathways.

Even the biggest AI systems today only have the rough digital equivalent of a few trillion “connections.” That’s impressive, but the brain is still vastly more flexible, creative and energy-efficient. A human brain runs on about 20 watts of power — roughly the same as a light bulb. Training a large AI model can use millions of times more energy.

What It Costs to Build AI

Creating frontier-level AI isn’t cheap.

• Training – The most advanced systems cost billions of dollars to train. In fact, it's estimated that training ChatGPT5 cost $10 Billion in electricity and hardware depreciation costs alone.d

• Hardware depreciation – Those chips and servers age fast. Every two or three years they need replacing, just like any other IT asset.

• People – Behind the code are thousands of engineers, researchers and data-labelers preparing the training material.

• Electricity & cooling – Large data centres use enormous amounts of power to keep everything running safely.

For small and medium-sized firms, building such models from scratch is unrealistic - and the cost of training AI models means it likely always will be. But using AI through cloud platforms or pre-built tools is now affordable — similar to paying for other software subscriptions.

The Future of AI in Everyday Business

Over the next few years, AI will quietly weave into almost every business process. A few examples:

• Customer service – Smarter chatbots and voice assistants that genuinely solve issues rather than frustrate.

• Admin automation – Drafting documents, replying to emails, scheduling, reporting.

• Marketing – Personalised content, SEO, and design produced in minutes.

• Cybersecurity – Predicting and blocking attacks faster than human analysts.

• Decision-making – Analysing trends and forecasting in ways Excel simply can’t.

AI won’t replace most jobs outright, at least not today, but it will change how work is done. The winners will be those who combine human judgment and AI speed effectively.

In Summary

The history of AI spans from 1950s theory to today’s practical tools. We’ve gone from researchers asking whether machines can think, to every business leader asking how AI can help them think faster.

The human brain still reigns supreme — but AI is catching up in narrow areas, and its cost of entry is falling. For SMEs, the opportunity lies not in competing with the tech giants, but in adapting faster: using AI to save time, cut waste, and deliver smarter service.