We are told that Artificial Intelligence is the new industrial revolution. Even conservative estimates compare it's impact to that of the birth of the internet. Hopes for efficient, cheap optimisation to modern economies, especially in the service sectors, has driven the stock market into an AI feeding frenzy. The controllers of modern western economies, the bankers, tech billionaires and neoliberal-minded politicians all have full-price, first class tickets on the Artificial Intelligence train, which is chugging away from the station, engine stoked with business to business marketing and stock market hype. The middle and working classes, however, are nervously stuck on a level crossing a few miles up ahead, hoping traffic clears before the train arrives.
While more realistic and socially grounded concerns, such as widespread cuts to entry level job markets, acceleration of already rampant consolidation of wealth, massive energy and non-renewable resource expenditure, are relatively well understood, they are often given equivalent weight to fanciful fears of a computational singularity, the domination of our species by an out of control thinking machine.
On the other hand, many see the advance of scientific and computational progress as an inevitability, a necessary step in the road to achieve a post-scarcity society. Utopian optimists, often the very same tech billionaires who sit comfortably in the dining car of the "AI train", are convinced that if we just build bigger and better thinking machines, we will be able to ask it to magically produce solutions to all of humanities big problems.
More realistic proponents simply think that the technology is useful, and it offers genuine value to businesses and the economy at large. Yes, many may find their entire field of work succumb to automation, they say, but that has happened before, during the industrial revolution and then in the widespread adoption of automated production lines. The economy will create other, more rewarding jobs for those people.
All of this noise about AI contributes to its image as the landmark technology of our age. The excitement and uncertainty, the billions being generated by exploding stock prices, the deluge of artificially generated text and images on social media, these things all make it very hard to dismiss AI's impact on conversations about the future of our societies and industries.
But at the core of this discourse, there is a real technology. That technology is specific, limited and ultimately, relatively simple. That this technology is not well understood by almost all of those who's lives it will impact, is a big problem. Understanding the basics of what a neural network actually is, and how it 'learns' or 'thinks', cuts away huge swaths of misinformation, enabling a clearer view of how AI is being used, and by whom, to benefit some at the expense of others.
I first intend to give you a tour of the locomotive of the AI train, and a peak under the hood if you are lucky. Then I'll show you the route of this routine service to Useful but Boring Technology Town, so that you can make your way down the carriages and tell the passengers that they'll have to change to a different platform if they have tickets to Utopia City or Doomsday Central. Maybe we can get station security to escort those salesmen selling all these mislabelled tickets off the premises.
But first, the fascinating topic of terminology.
