In 1997, IBM’s Deep Blue chess computer beat grand master Garry Kasparov. In 2011, IBM’s Watson won the US TV quiz show Jeopardy. Many experts now predict machine learning and artificial neural networks will form the next wave of innovation and disruption in the tech industry.
Giving a computer the capacity to reason, think and learn is almost as old as computing itself. Dubbed artificial intelligence (AI), Alan Turing formally presented his ideas on the subject as early as 1950.
Tech giants leading AI research
Fast-forward to now. The US Department of Defense was an earlier supporter of AI research and Australian institutions such as the DSTO, NICTA and CSIRO have been involved as well. Equally, major investors in AI today are the world’s biggest tech companies such as IBM, Chinese search giant Baidu, Google, Facebook, Apple, Microsoft and others.
While it consists of a variety of sub-disciplines, including speech recognition, knowledge representation and reasoning and computer vision, the basis of machine learning is that by feeding big data sets – the bigger the better – into algorithms, computers or robots are able to recognise patterns and try to predict future outcomes.
By then observing whether that prediction was correct or not, they are able to train themselves on how to best respond in certain situations. That’s a step beyond machine learning known as deep learning. The step beyond deep learning is even more mind-bending, with scientists and engineers building deep neural networks that seek to replicate the structure of the human brain.
The future of technology and business in an AI world
So what does this growing trend mean for the tech industry and Australian businesses? How will this influence the way we interact and work with technology?
For lovers of science fiction, the vision of conversing with AI seems tantalisingly close, but how will it really play out? Speech recognition continues to improve with every software upgrade and driverless cars are clearly going to happen as well, although not as soon as we may think.
Simplifying the day-to-day
In the short to medium term, machine learning promises to lighten much of the burden of day-to-day business administration. We’re probably not that far from supply chains that are able to maintain stock levels free of human intervention. The sub-discipline of computer vision is dedicated to developing the ability to filter out extraneous visual information and alert humans to the important exception. Picking out the face of a wanted criminal in a crowd or a downed airliner in the sea are typical examples.
Similarly, machine learning could be applied to much of the day-to-day work of IT. An AI could use machine learning to manage network security, doing the vital but at times boring job of monitoring all the various alarms to isolate the serious hacking attempt from the routine. In the short term, it would do what all task automation tends to promise: free up humans to do the higher-level, more strategic work.
A threat to our existence?
However, in the medium to long term, AI could represent an existential threat to humanity, with luminaries such as Steve Wozniak, Stephen Hawking, Elon Musk and Bill Gates all warning that when machines can autonomously design and make other machines, almost all of us could be out of a job.
This rather dystopian view has been developed at length by author Martin Ford in Rise of the Robots: Technology and the Threat of a Jobless Future, while people like Erik Brynjolfsson and Andrew McAfee have a more optimistic view in The Second Machine Age.
As the old Chinese curse goes: “May you live in interesting times.”