What history really tells us about AI’s potential

History is being used to argue both for and against AI, with citations of previous cultural events and practices as evidence. Unfortunately, however, most of the historical references used are just plain wrong. Regular readers will know that, after music, my biggest passion is history. I have a decent depth of knowledge of both. So, I’ve decided to use this knowledge to explain what history teaches us about how AI is likely to impact human creative endeavour.

Creativity and innovation benefit from mass scale

Firstly, let’s take the critique that AI’s mass creation will be a detriment to creativity and that scarcity is crucial. The most impactful shifts in human technology happened when scarcity was transformed into ubiquity. Here are three key examples:

1. Roman industrialisation

From the 1st century BCE to end of the 4th century CE, Roman industrialisation rocketed human invention and innovation forward. The Romans used advanced engineering – from iron-reinforced concrete in the baths of Caracalla to water-powered mills and hydraulic machines. Meanwhile, recent marine archaeology even suggests Roman galleys had magnetic compasses – centuries before such technologies reappeared in Europe. They also understood that the world was not flat, a fact not reestablished until the Renaissance. Roman industrial output was so vast that the ice rings show output did not reach comparable levels until the industrial revolution.

2. The Industrial Revolution

The Industrial Revolution transformed the Western world from a slow-moving agrarian economy into one defined by rapid innovation. It laid the foundations for today’s modern world. Because it created an educated middle class, education and learning accelerated, resulting in rapid advances in medicine, science, and technology.

3. The growth of universities

Since the post-war era, universities have become innovation and invention powerhouses. By pooling human talent, the rate of change and progress has advanced at unprecedented rates. As an example, look at the advances in astronomy in the last 20 years. The first exoplanet was discovered in 1995. As of 2025, 6,000 have been discovered (according to NASA).

History is awash with – and is defined by – examples of replacing scarcity with scale to drive progress. If Leonardo da Vinci had been part of a well-funded university, we could had human flight hundreds of years ago instead of the concept being a few drawings in one ‘scarce’ intellect’s notebook.

Technology accelerates creative endeavour

A common critique of gen AI is that it’s not creation – that a text prompt does not represent creative intent. This is a very narrow argument. Just because it’s being used by novice consumers to create so-called ‘AI slop’, does not mean that is what AI is or will be. AI is a technology with potential to supercharge human creativity. We have seen shifts before where technology once criticised by traditionalists went on to improve creative endeavour. Recent history is full of examples. Here are two:

1. Design software

When desktop design software began to take off, graphic designers said it was going to destroy human creativity. There is not much in the world that is designed without software these days.

2. Music software

When digital audio workstations (DAWs) and samplers started out, the traditionalist consensus was that this was going to result in synthetic music that was not true human artistry. Now there’s not much music in the world that doesn’t utilise a DAW, and most popular tracks include samples of some kind.

Ask what you can do for AI – not what AI can do for you

AI is a tool – it does what it is told to do. If you tell it to be creative and provide creative inputs, it will be creative. Are there ‘bad’ users of gen AI? Of course – just as there are ‘bad’ users of DAWs and graphic design software. If we were to eradicate technologies due to bad actors, then the Internet should be taken down and every car taken off the roads because of bad drivers.

It’s crucial to have checks and balances in place that, as far as is technologically possible, protect rightsholders and creators. I emphasise the ‘possible’ because we are fast approaching a point where it’s becoming hard for tech and licensing to keep up – not least as we move towards synthetic datasets (i.e., when AI starts learning from AI output, rather than human works). Giving creators the right to opt out of AI learning is a logical and ‘fair’ step to take. But, again, as we move towards synthetic datasets, even this will risk becoming ineffective.

Perhaps what makes AI distinct from DAWs and design software is that it is both a creative tool and a business model. Today’s rightsholders are, understandably, focused on protecting today’s revenue streams. However, the likelihood is that many gen AI companies will become destinations, where creation and consumption will become an unbroken continuum. People will turn to creation as entertainment and to entertainment as creation. The odds are that many gen AI companies will have little-to-no exposure to rightsholder licenses in this future world. So, creators and rightsholders will need to carve out new roles. Now is the time to work out what those roles are.

If history shows us anything, it is that when something new comes, it upturns the existing order. But humans work out how to turn it to their advantage. As the famous George Santayan quote goes: “Those who cannot remember the past are condemned to repeat it”.

Image credits: Joshua Humpfer /  Patrick Hendry / Solen Feyissa via Unsplash