The Machine is Learning – The Future is Here

Should we all be worried?

We live in an era of unprecedented change; and technology is trailblazing through all aspects of our lives. The doomsday scenario plays out that soon we will all be redundant, our jobs replaced by intelligent machines that work harder, faster and more accurately than we could ever do! Machines won’t just be tools to help us work smarter, they will become autonomous entities taking over ALL we do. Before long they have taken over and those of us who have survived form a resistance movement and the Terminator movie plays itself out? Recently 4 ‘Artificial Intelligence’ experts testified to the UK parliament that we are sleepwalking into this scenario; predictions include that we are only 45 years away from the moment of ‘singularity’. This is the point that a computer can improve itself; and indeed build new machines cleverer than itself and certainly smarter than us. That is an extreme prediction; more generally all 4 were in consensus that in the next 5 years machines would take over a swathe of tasks currently carried out by humans across multiple industries; the question is: so what?


According to research from Deloitte and Oxford University, Gartner and others 35% of current jobs are at risk from machines in the next decade. These aren’t all manual jobs either but white collar roles such as reporters, marketers, surgeons, financial analysts and many in the legal and insurance profession. This doesn’t even take into account the driverless car coming of age and all those Uber drivers and ultimately truck drivers having no job. McKinsey Global Institute recently published a more detailed piece of research whereby they examined a set of activities carried out by 750 different occupations. Over 2000+ activities were analysed and they concluded that 45% of these could be automated by technology that exists today. A further 13% of these could be automated by a fairly small improvement in the natural language understood by computers.

“When we modeled the potential of automation to transform business processes across several industries, we found that the benefits (ranging from increased output to higher quality and improved reliability, as well as the potential to perform some tasks at superhuman levels) typically are between three- and ten-times the cost,”

The conclusion is that not only ‘could’ the activity be taken over by a machine learning application but the machine would do a better job, faster, more reliable – delivering extraordinary benefits over a human.

Insurance Sector

McKinsey present detailed and interactive analysis of all the roles and activities for anyone to check how at risk their job is to machine takeover. The analysis of the insurance sector caught our interest; as it’s an area we at Chatterbox Labs are increasingly being asked by partners to create smart, data driven machine learning products for using our NLP Framework. By role they present the % of tasks currently carried out by humans that could be automated by machines; across the key roles there are significant amounts of automation to be delivered.

The report doesn’t focus on the exact tasks or activities to be automated per role; that will be expanded upon this year when they publish the full report.

Are all these people now redundant?

So ‘activities’ are at risk; ‘activities’ could be delivered faster, better, cheaper by machines; but critically the research from both sides of the Atlantic doesn’t conclude that people would be redundant. Even the Insurance Sales Agent with a potential of 60% activity automation isn’t replaced; there is still 40% of the role that at present can’t be automated. Rather than the machine taking over; the machine just does a much better job at specific things that perhaps humans shouldn’t be doing? The machine just takes away something we used to do allowing us to come up with something better to do? The argument goes that instead of making us redundant the machine is simply helping us to do what we have been doing for years – to advance.

Job Creation

Research from Deloitte articulates that “technology has created more jobs than its destroyed in the last 144 years.” Technology replaces the things we don’t really like doing; and only does so when it can do it faster, better and cheaper. The Deloitte research finds that machines have been “saving us from dull, repetitive and dangerous work” for decades and this is likely to be the trend moving forwards even though it appears these ‘smarter’ machines are taking over more ‘intelligent’ tasks. Technology in addition creates efficiencies in the economy; lowering the price of essentials which in turn provides for more disposable income which creates new demands and therefore jobs. It appears to be part of the rich progression of humanity as a whole; we adapt and change and keep creating and innovating when freed from the menial tasks of our ancestors. An estimate that the average American today is 80 times richer than their ancestor in 1900 would evidence the fact that progress has been constant and accelerating.

Smart Machines Won’t Kill Us

Machine learning technology that understands natural language is the future and it’s a good thing. Ultimately the technology will do the job better than humans, making the systems more efficient and productive; allowing us to focus on new innovations and ideas for industries that will create wealth and prosperity. Machine learning technology offers the opportunity to take away the mundane, driving us forwards to a better future. Whether the point of ‘singularity’ ever arrives; and indeed means the machines turn on us is highly unlikely (but we should keep an eye on it!).

Chatterbox Labs NLP Framework

Our extensible NLP Framework is used by partners globally to create disruptive, smart data-driven machine learning products that automate many tasks currently being either done manually or by software that, unlike ours, isn’t clever. Our statistical approach to machine learning provides for constant iteration of accuracy levels as the model understands what it is being asked to ‘interpret’.

If you have a specific use case for any industry or activity please feel free to get in contact with our VP Strategic Alliances, Andrew Watson, or view our web site for more information