The singularity and the future of intelligence
Everywhere you turn there is a discussion about the impact of technology on our future, whether it be how we work or how we live. Of particular concern is the encroachment of automation into virtually every part of our world.
Estimates vary, but it is credible to suggest that about half of all white collar work will be automated in the coming decades. However, the real revolution is often termed the “singularity”, in effect, the point where computers are smarter than we are.
Way back in 1962, John F. Kennedy said “I regard it as the major domestic challenge of the ’60s to maintain full employment at a time when automation is replacing men”. With just a little rewording, the same statement could be made today. We learn two things from this, the first is that the threat of automation is not new, the second is that properly managed there is no reason for the total number of jobs to decrease.
There will, however, be disruption which has social and economic implications. The jobs that will be created haven’t even been invented yet and not everyone working today will be equipped to take them up. Already we have worldwide and local shortages of technology workers while other sectors are shedding staff.
As machines get smarter, our role with them will become more complex. The trouble with the term smarter is that is that it conjures up images of a machine that is equivalent to us. As I’ve previously written, the limit of current technology is the ability to make “cognitive leaps”, that is the ability to determine something new from available data rather than repackage up an association that has been previously made. This boundary will protect many workers (see Your insight might protect your job).
However, as our technologies advance, the day is getting closer when we have to seriously think about what intelligence really means. The problem is we don’t know how the human brain achieves its feat of consciousness or how we achieve those cognitive leaps that are central to our intelligence.
Eminent quantum physicist Nicolas Gisin argues that free will and a conventional view of physics are in conflict (see New Scientist: Physics killed free will and time’s flow. We need them back). Gisin is well known for having successfully applied quantum mechanics to create commercial applications and also to have demonstrated some of the more controversial conclusions of the field such as quantum teleportation.
Gisin asks “Are we passive laundry machines through which thoughts happen to pass? Or are we active agents free to influence our thoughts and decisions?” He goes on to argue that modern science tends towards the machine view, he says “In a deterministic universe, where one thing leads inevitably to the next, any conception we have of free will is an illusion.”
Meanwhile, another physicist, Matthew Fisher, has speculated that our brains probably rely on quantum mechanics to achieve the magic of consciousness (see New Scientist: Is quantum physics behind your brain’s ability to think?).
Until now, the objectives of the many teams racing to develop quantum computers have been to solve algorithms that are in the domain of conventional computers but too ambitious for today’s processors. I’ve argued that these projects may not live up to their promise (see The Quantum Computer dream could be killed by information management).
It is possible that the really interesting outcome from current research into Quantum Computing could be a machine that displays many of the same qualities as our own brain. Until the workings of such a machine is designed, any date for the “singularity” is nothing more than speculation.
Of course, when we do design such a machine, there will be some very difficult ethical and social considerations. I’ve argued before that despite the threat that such machines pose, ultimately Our machines won’t outsmart us, mainly because we are tough evolutionary creatures and we will assimilate the technology before we will let it rule us.
Regardless of when we manage to invent machines that can take the next step in artificial intelligence, there is no doubt that to both live and work in the twenty first century we need to be prepared to adapt, and quickly.