That was the question posed to a packed room at last night’s Executive Discussion Evening – a big question, yes, but a relevant one too.
You don’t have to search for long to find media reports or expert commentary suggesting that advances in technology will soon bring the world to its knees.
Robots driven by artificial intelligence (AI) rising up to overcome mankind has been a common theme of science fiction for years.
In fact, one of last night’s panellists – Murray Shanahan, Professor of Cognitive Robotics at Imperial College London – served as an advisor to the team behind the 2015 sci-fi hit Ex Machina (a film which, you guessed it, tackles the topic of humanising AI).
Today we find ourselves living in a world that is increasingly automated, and in some cases becoming dependent on automation. But are we all going to go into work tomorrow and find we’ve been replaced by a robot? Probably not, says Shanahan.
Developments in AI, he told the room, have been driven by considerable success in the sub-fields of machine learning and computer vision. Just 10 years ago, the technology we treat as commonplace today would have been well beyond possibility.
Increased computing power (particularly within graphics processing units, which have been co-opted from high-end video games to help drive machine learning algorithms), as well as the boom of big data and internet of things applications have helped usher in this new reality.
Thanks to significant corporate investment and a buoyant start-up scene, we can expect to see very significant disruption caused by AI within the next 10 years.
So will we all be serving our robot overlords in 10 years’ time? Again, probably not.
AlphaGo – Google DeepMind’s AI program designed to play the ancient board game Go – achieved something remarkable this year when it beat 18-time Go world champion Lee Sedol in a five-game match.
This was particularly remarkable, Shanahan explained, because the program had to employ a much more human approach (i.e. one which resembled human intelligence) to beat Sedol. The victory involved not only a deep understanding of the complexity of the game, but also much less tangible traits such as intuitive understanding and creativity.
As impressive as this is, AlphaGo represents the contrast between the short-term and long-term challenges of AI.
Most of today’s examples of AI, and those which pose a genuine threat to both service and knowledge workers in terms of job security, Shanahan explained as specialist AI technology.
AlphaGo, he said, is an example of state of the art AI and is indeed quite clever. But ultimately it can’t do anything except play Go. Unlike Lee Sedol, it can’t hold a conversation, play with a child or make a meal – these things would require what Shanahan called AGI, or artificial general intelligence.
We don’t yet know how to create this. As complex as a game of Go is, everyday life is much more so – indeed, everyday life includes Go boards…
The future is a long way away
There’s little doubt that specialist AI is going to have a big impact over the coming decade, but a world run by programs executing human intelligence is a long way off, despite what you might read in the papers.
Rough calculations suggest that computing power that can match that of the human brain would require Exaflop computing (which involves 1018 floating point operations per second). At today’s exponential development rate, it will be 2025 before that is possible, and then we’ll have the challenge of working out how we actually use that power to create AGI.
Hamish McRae, a forecaster, author and journalist – whose job could be described as predicting what’s going to happen in the future (Lottery numbers not included) – joined Professor Shanahan to ask how such technology can be used to benefit society.
There’s nothing new, he said, about worrying that the world is going to hell in a handcart. People as far back as Confucius and Malthus have been predicting it at almost every opportunity.
That’s not to say that we as a human population don’t face a great challenge but, he said, echoing the American author Joseph Wood Krutch, while technology made large populations possible, large populations have now made technology indispensable.
Think of the children
How then do we tackle the problem of ageing populations in the developed world, and ensure an increase in living standards that will benefit our children?
Hal Varian, Google’s chief economist, has argued that there isn’t so much a problem with the stagnation of living standards today, but rather a problem with how we’re measuring living standards.
He argues that many of the things that improve our lives today are free, so we can’t apply the same measurement technique as we previously have done in the past. Unless, of course, you want to see this as an extension of the aphorism that the best things in the life are free.
The growth of mobile communications – and social media with it – has been instrumental in both a qualitative rise in living quality and an ongoing shift of power to big-population countries.
With technology functioning as a democratiser, and smartphones, social media and messaging services changing the world, McRae suggested, the challenge then is of how to maximise social capital. Essentially, how can we adapt to these changes to create a society that works for everyone and not just the few.
AI is bound to play a vital role in this, not only in the threat it poses to both service and knowledge industry jobs but also in helping us adapt to this new world and create new opportunities.
Key to this will be using AI and related technologies to make the service industries more productive, in the same way that we’ve achieved with good industries over the past years.
To do this, McRae said, we will need to embrace our fears and put some faith in the strange fact that humans have a habit of accidentally getting things right – not so much as a result of what we know but, rather what we don’t know will happen.
Perhaps it is just this accidental tendency – you might call it instinct – that sets us apart from the machines.
After all, McRae concluded with a quote from Albert Einstein, “information isn’t knowledge.”
Continue the conversation online using the #FujitsuEDE hashtag, or by joining the EDE LinkedIn group through which you can receive advance notice of upcoming events.
The next EDE event will take place on 1st March 2017. For more information on other upcoming events please click here.
Tom is Fujitsu’s Executive sponsor for our Graduates and Apprentices and is a fully accredited and qualified Executive Coach.
Latest posts by Tom Roche (see all)
- Walking with robots: is this the end of the world as we know it? - October 7, 2016