Robert Lea, Head of Global Equity Research, discusses the rise in artificial intelligence, robotics and automation. Robert has closely followed developments in AI for more than 20 years, both as a technology strategist and electronics engineer.
Recent rapid advances in technology have led to a surge in media interest in artificial intelligence (AI), robotics and automation. While some news stories have focussed on the more positive aspects of these new technologies, such as vehicle automation, others have chosen to take a more negative slant.
For example, a recent PwC report suggested up to 30% of UK jobs were at risk from automation, by 2030. A number of high profile figures from Bill Gates, to Elon Musk and Stephen Hawking have also taken to the public platform to warn about the longer-term potential dangers of AI.
So where does the truth lie? Should we be fearful of the future, or embrace it? We are more inclined to believe the latter, and expect these evolving technologies to ultimately make a positive contribution to society and the broader economy as a whole.
Take AI, for example, which is already widely used within smartphones and social media. AI algorithms are well suited to pattern recognition, which makes them ideal at tagging photos on Facebook, or tailoring newsfeeds and advertising, based on a user’s past history. AI computational engines also provide the processing power behind the Cortana and Siri Digital Assistants, as well as forming an integral part of online search engines, such as Google Chrome.
While AI is adept at analysing data, the current generation of AI solutions are application specific and are only able to operate within the narrow confines of the task they were designed for. For example, the Cortana digital assistant would have zero chance of being able to recognise a face from a photograph, as it was not built for that role.
Critics of this piece will argue that AI does have the ability to learn. While this is technically true, the learning ability of any AI program is still confined to the function it was designed for. The term ‘learn’ is also potentially misleading, as it really describes the ability of AI networks to ‘learn’ and improve their performance through the repeated performance of the set task. The ability of AI systems to learn outside the confines of their programmed function is extremely limited.
In other words, AI - as it exists today - is not really very intelligent at all. AI systems struggle when it comes to creative tasks, let alone being able to think for themselves. As a result, the use of the word ‘Intelligence’ is a potential misnomer, as current AI systems do not possess an intelligence in any conventional sense.
Where AI does pose a theoretical risk, is where it has access to, or has control over a safety critical system, e.g. operating the control rods on a nuclear reactor, or driving an unmanned road vehicle. However, this risk can be mitigated by incorporating sufficiently robust redundancy into the system.
So should we be worried about the rise of the Machines? On balance no, although rapid progress has been made in recent years, AI developments still remain at a relatively early stage. As things stand, the prospect of a sentient machine, capable of original thought and possessing an artificial consciousness remains the stuff of Science Fiction. This is not to suggest this will never happen, but it is not a breakthrough I expect to occur in my lifetime.
While we might not need worry about the near-term threat from AI, should we be concerned that robots are about to replace our jobs? Again, we believe the threat from robotics and automation has been over-stated in the press and see developments in these areas as more of an opportunity, than a threat.
First, consider that the International Federation of Robotics estimates less than 10% of current jobs globally are fully automatable. This number flies in the face of recent, more alarmist reports and is all the more noteworthy given it comes from the Robotics industry trade body – an organisation that has every incentive to promote the benefits of automation.
Second, and perhaps unsurprisingly, a large proportion of the at-risk roles are in manufacturing, which forms a relatively small part of the UK economy. The UK manufacturing sector – dominated by automotive, aerospace and pharmaceutical sectors – is already highly automated. As such, suggestions that 30% of UK jobs are at risk from automation appears somewhat wide of the mark, in our view.
On balance, we see the increasing use of robotic and automation as a net positive, and expect them to raise productivity in the economy, rather than replace jobs directly. In other words, the use of robotics and automation should augment labour, not substitute it.
This is not to suggest that the jobs of some lower-skilled workers will be at risk in the future. They will. But this is nothing new, and is a trend that has been in play since the dawn of the industrial revolution. It is therefore incumbent upon Governments help the affected workers re-train and invest in developing new skills.
The rising use of technology should also be seen in the context of the demographic challenges faced by many countries – namely, an ageing population. Robotics and Automation can help plug the skills gap left by a shrinking working population. Robotics also has a role to play in caring for the elderly, an important benefit given a child born in 2017 Britain is expected to live into their 90s.
So we should embrace these new technologies, not fear them. When used appropriately, AI, robotics and automation can make a positive contribution to society, raising productivity in the economy and ultimately enhancing people’s lives.