Self-Driving Cars, Robots – and Paperclips? Artificial Intelligence and its Implications

Date
Author:
Amy Corkery, NCPO

Universal Paperclips (http://www.decisionproblem.com/paperclips/index2.html) is a new game sweeping the world’s offices, and to describe it as ‘addictive’ would be an understatement. Consider yourself warned.

You start by clicking a button that makes a paperclip. You can spend money on marketing, adjust the price, and buy more raw materials periodically. Sounds reasonably like enterprise in the ‘real world’, right?

Well, sort of. A little while into playing the game, you get “computational resources” and can undertake ‘projects’ to boost paper-clip productivity. Then you start to feel your role, as the human in control, taking a turn.  The projects change from things like “Improved Auto-Clippers” to things like “Quantum Foam Annealment”. Everything takes care of itself without any input from you for five minutes, then fifteen minutes, then an hour. At a certain point, you feel the familiar buy-sell-invest trappings of a business give way to you making the occasional weighty decision (‘do I solve global warming, or male pattern baldness?’) and otherwise letting the computers do all the work. 

Without giving too much away, you’re then left with entirely more paperclips than anybody would ever want.

Universal Paperclips carries food for thought, regarding a future where computers do more and more of the ‘thinking’ for us – what’s known as artificial intelligence (‘AI’). It’s not quite a full or accurate depiction (well, hopefully we won’t see businesses deploying drones to hypnotise customers into buying their products!) – but it is a profound look at the collision of capitalism, economics and AI: in short, a feature of the future of the way we work.

We’re seeing the start of this now: artificial intelligence is used in smartphone voice assistants, online video recommendations, and ridesharing apps. AI is also being used in research and information processing; and in sales, consumer experience, and customer assistance. 

We can’t know exactly where to next, but we can guess. Along with the tech industry, AI disruption is likely in the legal, finance, agriculture, transportation, medicine, and service industries. AI doctors, AI sprinkler systems, AI disaster prediction detectors, and AI home helpers are all within contemplation. The way we work might fundamentally change: replacing ‘humdrum’ ways of working with more ‘creative’ ones.

Of course, there are some possible downsides and philosophical quandaries here. These include

- safety and security concerns (what happens if an AI is programmed to do something devastating, is compromised, or uses a negative method to achieve a positive goal?);

- privacy and civil liberties concerns (how much of people’s lives will the government and private companies know about?); and

- the politics of what will happen when a large amount of jobs become obsolete as a result of AI.

There are exciting times and big changes ahead, and a lot of opportunity for humanity as a whole. And we know there will definitely be enough paperclips to go around.