Why everyone needs to be involved in AI

To look into the future, it is first wise to look into the past. Human society has already been through at least two major revolutions, both of which have dramatically changed the way we live.

First there was the agricultural revolution. Before that humans didn’t normally live in fixed places, but moved around to maximise their hunting and gathering potential. The invention of agriculture bought planting and harvesting crops, which in turn bought farms. People stayed put and held onto the best land, passing it down to their children. Farms could supply villages where everybody didn’t have to grow or hunt their own food – they could specialise into making things such as shoes and farm implements and art which they could trade. The people left hunting and gathering could not compete.

The next major change in human society came with the industrial revolution. Factories needed labour, and labour meant towns and cities with suburbs. Material good were produced faster and cheaper and to higher quality than ever before. The industrial revolution took us into space, and for the most part laid the foundation for the world we have today.

We are currently at the start of the digital revolution, and the impacts on our society are going to be just as great as the previous revolutions. The physical world has begun to be mirrored and influenced by an invisible digital domain. For example, without most of us noticing, our money changed from being intrinsically worth something (being linked to gold) to being binary digits in computer storage, with no intrinsic worth. Algorithms started to trade in financial markets, creating and destroying the value of real world companies (see the flash crash of 2010). On the high street, well established retailers who were tied into expensive in-town buildings went out of business to retailers who sold on-line and could minimise their costs. The richest people in the world ceased being those related to the production of physical stuff, and started to become those related to the utilisation of data. And yet we haven’t really started.

We are at a point where a number of key technologies have together reached enough capability to trigger massive change in the way we live. The internet, 3D printing and AI are three of the most visible, and of these AI has the most potential for disruptive change. Let’s leave aside Hollywood style AI threats from HAL or Skynet and look at the immediate real world. AI, by its nature, is attempting to automate the activities that the machines of the industrial age could not do. The industrial revolution may have taken away the jobs of many village shoe makers and blacksmiths, but AI will quickly start to take away or transform the jobs of the current middle class: lawyers, solicitors, doctors, artists.
So where are we heading? At the moment AIs tend to be specialised for specific tasks – piloting a car, producing a medical diagnosis or composing music. But we’ve seen this before. We used to have sports cars for weekend jaunts down country lanes and more domestically friendly cars for carrying the kids and the dog to the shops, now you can buy an estate car which will break 60 miles an hour from a standing start in under 3.5 seconds (not recommended if you have the kids and a dog in the back at the time). I used to own a camera, a calculator and a GPS device… you get the idea. More general purpose AIs are inevitable; true, it won’t be easy – but that just makes it more attractive as a goal.
Currently we are exploring the intellectual limits of what we can achieve, however given the potential impacts on our society, I think that we need to start to make more serious efforts to create a strategic view. I also think that it is important that the view should as far as possible be a shared one. AI is something that everyone needs to be familiar with, at least enough to be able to take part in important discussions; Should AIs operate under a set of rules (e.g. Azimov’s three rules of robotics)?; Should we allow AIs to have emotional states and drives of their own?; What, if any, limits should we place on AIs making significant decisions which affect the direction of our society?
Uniquely, I think we are in a position to finesse the outcome of this digital revolution. Because the impacts are going to be larger, more far reaching and on a greater population of people than ever before, I think we are beholden to try our best.