"Shall We Play a Game ?" The Month in Technology Mar 2016, Jersey Evening Post

This month I return to a topic that I touched on previously, Artificial Intelligence (“AI”).

Recently I gave a talk to the Highlands business breakfast discussing how technology is going to affect jobs, and the education our next generation will need to prepare them for this new World.  I had written about this last year, but thought I would check recent information.

I was amazed by how much has changed, particularly that the estimate that 25% of US jobs will be automated in the next 20 years has since been updated to 47%.  In China that is 77% !

In the UK RBS recently announced that they will be using so called computer “robo advisers” to provide investment advice to anyone with assets under £250k, in the process cutting 550 wealth management jobs.

More widely reported, the British designed Google Deep Mind last week beat a world expert at the Chinese game of Go.

It’s interesting that a computer can beat a human at a game, but relies on a human to move the pieces and doesn’t actually realise it is playing a game at all just running through huge combinations of potential moves.

This shows the difference between pattern identification, which computers are very good at and getting better, and true consciousness.  There is much speculation about how long it’s going to take for computers to become aware of their own existence, what’s known as “The Singularity” and a “Takeoff”.

At this point computers will be able to redesign and build themselves, and is estimated to happen in the period of year 2029 to 2045.

From this point on the possibilities and dangers put forward by scientists and futurologists feel more like movie summaries, reflecting the theory that horror film plots reflect society’s fears.

Risk - Private corporations may not share their AI for commercial reasons, causing great risk to us all – see Terminator and Robocop.

Risk - One of the first potential uses for sentient AI will be space exploration, but we have to be careful about them coming back to Earth smarter than us – see Blade Runner.

Risk - A bug in AI code may lead computers to break Asimov’s laws and kill humans – see lots of films including 2001, I Robot, War Games and Dr Strangelove.

The doomsday scenario is that AI outstrips our capabilities to the extent that they will only need us as amusing pets reflecting their organic origins, much as we feel when we see Guernseymen.

There are several startups in Jersey that will be building AI solutions, and whilst we don’t need to immediately fear time travelling robots it’s vital that conversations take place about how AI will be controlled in the human interest.